Mixed donor chimerism under three years of age - the same drug exposure ablates less marrow in a younger child, and dosing cannot fix it
also called mixed chimerism · donor myeloid chimerism · DMC · incomplete donor engraftment · minimal donor chimerism · age effect on ablation
After a transplant, doctors check what fraction of the child's blood cells come from the donor. In children under three, nearly half end up with a mixture of their own and the donor's cells, while in older children almost none do - and this happens even when they received exactly the same amount of the conditioning drug. In a leukaemia, the cells that were not replaced are where a relapse comes from.
The same thing, for a scientist
Donor myeloid chimerism in peripheral blood is the standard surrogate for marrow haematopoietic stem cell replacement. In 105 children with non-malignant disease conditioned with myeloablative busulfan (cumulative AUC at or above 55 mg*h/L), the 3-year cumulative incidence of mixed donor myeloid chimerism (below 95%) was 45.2% in those under 3 years against 1.9% at 3 years or older (p<0.001), and minimal chimerism (20% or below) was 12.8% against 0% (p=0.008) - with no difference in busulfan exposure between the age groups [E1]. Busulfan exposure did also matter, with mixed chimerism in 24.8% at cumulative AUC 55-74.9 mg*h/L versus 5.0% at 75 mg*h/L or above, but in children aged 3 years or older exposure had no effect at all [E1]. The authors attribute the age effect to greater marrow reserve in younger children, and note the same phenomenon with melphalan-based conditioning. This is a pharmacodynamic age effect, not a pharmacokinetic one: perfecting exposure-guided dosing would not have closed the gap.
The finding, and why the second clause is the important one1 deeper
| outcome at 3 years | age <3 years | age >=3 years | p |
|---|---|---|---|
| mixed donor myeloid chimerism (<95%) | 45.2% (29.7-60.7) | 1.9% (0.1-5.4) | <0.001 |
| minimal donor myeloid chimerism (<=20%) | 12.8% (2.2-23.4) | 0% (0-6.1) | 0.008 |
And: 'there was no difference in busulfan exposure between the groups' 1.
Everything in this beat about dosing units, therapeutic drug monitoring and exposure targeting concerns getting the right amount of drug into a small patient. This finding is orthogonal to all of it. These children received the same exposure and the drug did less.
The proposed mechanism is greater marrow reserve: a younger child has more haematopoietic stem cells per kilogram to ablate, so the same alkylating exposure clears a smaller fraction of them. It is a hypothesis in the paper and was not measured. If it is right, the implication is that infant conditioning may need to be more intense per kilogram than older-child conditioning - the opposite direction from every age-based dose reduction in the frontline protocol.
Exposure still matters, but only in the younger group
Within the same cohort 1:
| cumulative busulfan AUC | mixed donor myeloid chimerism at 3 years |
|---|---|
| 55-74.9 mg*h/L | 24.8% (14.8-34.8) |
| >= 75 mg*h/L | 5.0% (0.1-8.3) |
And 'in patients >=3 years of age, there was no impact of busulfan exposure on development of mixed DMC'.
So the exposure-response relationship for marrow replacement exists only in the young. In older children the ablation is complete across the studied range and exposure is irrelevant to this endpoint. That is exactly the signature of a reserve effect: where the target pool is large relative to the dose, dose matters; where it is small, it does not.
Note the lower bound. The FORUM optimal window starts at 73.3 mg*h/L and Bartelink's at 78; the chimerism threshold here is 75. The exposure that produces complete donor replacement in a young child is the same exposure the leukaemia literature independently identifies as the floor for avoiding relapse.
Why this matters more in leukaemia than in the disease it was measured in
The cohort is non-malignant disease, where mixed chimerism means partial correction of a genetic defect and is often clinically acceptable - for many conditions 20-30% donor myeloid chimerism is curative.
In leukaemia it means something else entirely. Residual host haematopoiesis is the compartment the leukaemia came from and the compartment a relapse will come from; falling donor chimerism is a recognised trigger for withdrawing immunosuppression or giving donor lymphocyte infusion. The same measurement is a partial success in one setting and an early warning in the other.
So the honest position is that this finding has not been demonstrated in infant KMT2A-rearranged ALL, and its translation is a matter of mechanism rather than of evidence: if the effect is marrow reserve, reserve does not know what disease the patient has, and an infant transplanted for leukaemia should show the same incomplete replacement. If it is something about non-malignant disease biology - less prior chemotherapy, an intact marrow niche, no prior marrow damage - it may not transfer at all. An infant with KMT2A-rearranged ALL reaching transplant has had months of intensive chemotherapy first, which is precisely the difference.
| kind | resistance |
|---|---|
| severity | 45.2% mixed donor myeloid chimerism at 3 years below age 3, versus 1.9% at or above age 3, at identical busulfan exposure; minimal chimerism 12.8% versus 0% |
| mitigation | higher busulfan exposure helps in the young (mixed chimerism 24.8% at cumulative AUC 55-74.9 mg*h/L versus 5.0% at >=75) but has no effect at all in children 3 years and older - so the only available lever operates precisely in the age group at risk |
| mechanism proposed | greater marrow haematopoietic stem cell reserve per kilogram in younger children, so the same alkylating exposure ablates a smaller fraction; the authors note the same phenomenon with melphalan-based conditioning. NOT measured. |
| why it matters differently in leukaemia | in non-malignant disease mixed chimerism is partial correction and is often acceptable or curative; in leukaemia residual host haematopoiesis is the compartment the disease arose from and falling donor chimerism is a recognised trigger for withdrawing immunosuppression or giving donor lymphocyte infusion |
| threshold juxtaposition | 75 mg*h/L (chimerism, non-malignant, under 3 years) sits 1.7 mg*h/L above 73.3 mg*h/L (FORUM relapse threshold, paediatric ALL) - suggestive of a common ablation requirement, not evidence of one |
| the direction it implies | if the reserve hypothesis is right, infant conditioning may need to be MORE intense per kilogram, not less - the opposite direction from every age-based dose reduction in the Interfant-06 chemotherapy chapters |
| untested in this disease | the cohort is non-malignant; an infant reaching transplant for KMT2A-rearranged ALL has had months of intensive chemotherapy first, which is exactly the variable that would change marrow reserve |
affects
[ "drug-busulfan", "drug-melphalan" ]
- This is the one age effect in the beat that dosing cannot fix. Everything else here is about delivering the right exposure to a small patient; this is about the same exposure doing less, which means perfect therapeutic drug monitoring would not have prevented it.
- The implied correction runs opposite to every other infant adjustment. If younger marrow needs more alkylator per kilogram to ablate, then the age-based dose reductions that are correct for frontline chemotherapy are exactly wrong for conditioning - and the protocol applies neither.
- Residual host haematopoiesis in a leukaemia is not a partial success. The same laboratory number that means a genetic disease is corrected means, in an infant with KMT2A-rearranged ALL, that the compartment the leukaemia came from was not replaced.
- The only lever that works is the one with the veno-occlusive ceiling above it. Raising busulfan exposure to 75 mg*h/L or more reduces mixed chimerism in the young, and the hepatotoxicity ceiling sits at roughly 98 - a usable window, but a narrow one to be navigating with optional monitoring.
- The finding comes from a population whose marrow was never damaged by chemotherapy, which is the variable most likely to make it not transfer - so an entry citing it as an infant leukaemia risk is one inferential step ahead of the evidence, and says so.
- Does the under-3 chimerism effect operate in infants transplanted for leukaemia, who have had months of intensive chemotherapy that would deplete the marrow reserve the mechanism depends on? The effect could be absent, or the chemotherapy could simply move the threshold.
- Is donor chimerism the mediator of the busulfan exposure-relapse relationship in paediatric ALL? The two thresholds sit 1.7 mg*h/L apart across independent studies, and testing mediation requires measuring chimerism and relapse in one leukaemia cohort - which no published study does.
- If younger children need more busulfan exposure for complete ablation, does the optimal window differ by age? FORUM's window of 73.3-98.0 mg*h/L was derived across 0.5-19.5 years without age stratification, so it may be a blend of two different windows.
- Is marrow reserve actually the mechanism? It is a hypothesis, it was not measured, and it is directly testable by CD34+ quantification before conditioning - which is routinely available.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In 105 consecutive children with non-malignant disease (median age 3.9 years) undergoing first allogeneic HCT with myeloablative busulfan-based conditioning (cumulative AUC at or above 55 mg*h/L) from 2007 to 2023, the 3-year cumulative incidence of mixed donor myeloid chimerism was 45.2% (95% CI 29.7-60.7) in those aged under 3 years versus 1.9% (0.1-5.4) at 3 years or older (p<0.001), with no difference in busulfan exposure between the groups; minimal donor chimerism was 12.8% versus 0% (p=0.008); busulfan exposure was associated with mixed chimerism overall (24.8% at cumulative AUC 55-74.9 versus 5.0% at 75 mg*h/L or above, p=0.03) but had no impact in patients aged 3 years or older; and there was no impact of age or exposure on graft rejection, acute or chronic GVHD, transplant-related mortality or sinusoidal obstruction syndrome.
Younger children with nonmalignant disease have increased incidence of mixed myeloid chimerism after allogeneic hematopoietic cell transplantation with busulfan-based conditioning. · Dvorak CC, Temple WC, Cisneros GS, Chu J, Winestone LE, Higham CS, Shimano KA, Kharbanda S, et al. · Transplantation and Cellular Therapy 2025What this evidence cannot show
Non-malignant disease only, retrospective, single institution, 105 patients accrued over 16 years during which conditioning practice changed. The age cut-off of 2.9 years was derived by receiver-operating-characteristic analysis within the same dataset rather than prespecified, so it is a fitted boundary and not an externally validated one. The comparison is above and below 3 years; whether the effect is larger below 1 year is not reported. The proposed mechanism of greater marrow reserve is the authors' hypothesis and was not measured. Mixed chimerism carries different clinical meaning in leukaemia than in genetic disease.
Evidence that argues against this (1)
- The age effect did not propagate to any clinical endpoint in the same study: no impact of age or busulfan exposure on immunologic graft rejection, acute or chronic GVHD, transplant-related mortality or sinusoidal obstruction syndrome. A 45.2% versus 1.9% difference in a surrogate that does not move any outcome it was tested against may be a laboratory phenomenon rather than a clinical one, and in non-malignant disease mixed chimerism is frequently the desired result rather than a failure. Dvorak CC et al. 2025
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In paediatric ALL within the FORUM trial, cumulative busulfan exposure below 73.3 mg*h/L was associated with higher relapse rates (HR 1.93, p=0.019), placing the underexposure boundary for relapse avoidance within 2 mg*h/L of the 75 mg*h/L exposure above which mixed donor myeloid chimerism largely disappears in young children.
Refining busulfan exposure enhances pediatric ALL HSCT outcomes: insights from the International FORUM study. · Ben Hassine K, Gloor Y, Dupanloup I, Uppugunduri CRS, Mlakar V, Gonzales F, Guengoer T, Ifversen M, et al. · Blood Advances 2026What this evidence cannot show
The juxtaposition of 73.3 mg*h/L (FORUM relapse threshold in paediatric ALL) and 75 mg*h/L (chimerism threshold in paediatric non-malignant disease) is this beat's observation across two independent studies with different populations, endpoints, conditioning regimens and analytical methods. Both thresholds were derived within their own datasets. The proximity is suggestive of a common underlying quantity - the exposure required for adequate host haematopoietic ablation - and it is not evidence of one. No study has tested whether donor chimerism mediates the busulfan exposure-relapse relationship in leukaemia.
Evidence that argues against this (1)
- Two thresholds landing 1.7 mg*h/L apart across unrelated studies is also exactly what would happen by chance given that both were fitted within their own data and both lie in the middle of the clinically used exposure range. Data-derived cut-points in overlapping ranges converge easily, and treating the coincidence as mechanistic would be the kind of inference this project's brief warns against for database facts. The mediation claim requires measuring chimerism and relapse in the same leukaemia cohort, which nobody has done. Bartelink IH et al. 2016
mixed chimerism 24.8% at cumulative AUC 55-74.9 mg*h/L versus 5.0% at >=75, in children under 3 only
a pharmacodynamic age effect at identical exposure - the one age difference dosing cannot correct
residual host haematopoiesis is the compartment a post-transplant relapse arises from
the same age effect on chimerism is reported for melphalan-based conditioning