Oncofusion

start / graph / drug-busulfan

Drugstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q53-transplant

Busulfan - the alkylating agent that replaces radiation in infant conditioning, and whose labelled paediatric target range is below the range the modern trial data support

also called busulfan · busulphan · BU · Bu · Busulfex (IV, US trade name) · Busilvex (IV, EU trade name) · Myleran (oral tablet trade name) · 1,4-butanediol dimethanesulfonate · bifunctional alkylsulfonate (class)

Busulfan is the chemotherapy drug used instead of whole-body radiation to empty a baby's bone marrow before a transplant, because radiation damages a growing brain and body. It is very hard to dose in a baby: too little and the leukaemia comes back, too much and the liver is injured, so the blood level has to be measured during treatment and the dose corrected.

The same thing, for a scientist

Busulfan is a bifunctional alkylsulfonate that cross-links DNA and is the myeloablative backbone of irradiation-free conditioning in infants, who are excluded from total body irradiation on developmental grounds. Its pharmacokinetics are highly variable and age-dependent, its therapeutic window is narrow, and exposure - not dose - is what predicts outcome. Interfant-06 uses oral busulfan 20 mg/kg total (1.25 mg/kg every 6 h for 4 days) with cyclophosphamide and melphalan, and makes therapeutic drug monitoring optional ('blood levels of the drug should be measured if possible') [E1]. The FORUM trial's pharmacokinetic substudy defines the optimal cumulative AUC as 73.3-98.0 mg*h/L, with underexposure driving relapse and overexposure driving toxicity and GVHD [E2]; a 674-patient multicentre cohort independently put the optimum at 78-101 mg*h/L [E3]. Converted into the same units, the paediatric target range on the US intravenous busulfan label (900-1350 uM*min per dose x 16 doses = 59-89 mg*h/L) has a lower bound roughly 14 mg*h/L BELOW the FORUM underexposure threshold [E4].

Go deeper
Why busulfan and not radiation, in a baby1 deeper

The Interfant-06 protocol states the reasoning in one sentence: 'Busulfan (BU) is an established myeloablative alternative to total body irradiation (TBI). An irradiation free conditioning with busulfan, melphalan and cyclophosphamide that has shown good results in JMML and AML, will be used.' And in the stratification chapter: 'Total body irradiation is not used because of its neurotoxic side effects in these young children.' 1

So busulfan's place in infant leukaemia is defined by what it is not. It is not chosen because it outperformed TBI in an infant trial - no such trial exists. It is chosen because TBI is judged unacceptable at this age, and busulfan is the best-characterised myeloablative alkylator available to fill the gap. That is a defensible clinical decision and a weak evidential position, and the protocol says so plainly.

The exact Interfant-06 regimen, as written

From the protocol's chapter 6.5 1:

DayAgentDoseBasis
-7 to -4Busulfan5 mg/kg/day, i.e. 1.25 mg/kg orally every 6 hbody weight
-3, -2Cyclophosphamide60 mg/kg i.v.body weight
-1Melphalan140 mg/m2 i.v.body surface area
0Allogeneic HSCT

For matched (non-sibling) donors, ATG-Fresenius 20 mg/kg is added on days -3, -2 and -1.

Total busulfan dose is 20 mg/kg, given orally. The protocol adds: 'Blood levels of the drug should be measured if possible, and the dose should be adjusted accordingly. During BU-application a seizure prophylaxis is required. If intravenous busulfan is applied the given preparation and dose must be documented.' Therapeutic drug monitoring is therefore permissive, not mandatory, and the route is not fixed - two facts that matter enormously given everything in the next section. See concept-weight-versus-bsa-dosing-in-infant-conditioning for the dosing-unit problem visible in that table.

Exposure, not dose, is what predicts outcome - and the numbers now disagree with the label

Two large datasets converge on an optimal cumulative exposure window and then differ at the edges:

sourcepopulationoptimal cumulative AUCwhat falls outside
FORUM PK substudy, n=145, ages 0.5-19.5 y 2paediatric ALL, Flu/Bu/Thio73.3-98.0 mg*h/L<73.3 -> relapse HR 1.93; >98.0 -> toxicity and GVHD
Multicentre cohort, n=674, median age 4.5 y 3malignant and non-malignant78-101 mg*h/L2-y EFS 77.0% optimum vs 66.1% at the low historical target
US i.v. busulfan label, paediatric PK study 4n=24, 5 months-16 years900-1350 uM*min per doseHVOD risk stated above 1,500 uM*min

The third row is in different units and that is the trap. Busulfan's molecular weight is 246.3, so 1 uM = 0.2463 mg/L and 1 uMmin = 0.2463/60 mgh/L. A per-dose AUC of 900-1350 uMmin over the standard 16 doses is therefore **59.1-88.7 mgh/L cumulative**. Its lower bound sits about 14 mg*h/L below FORUM's underexposure threshold, inside the region FORUM associates with a near-doubling of relapse hazard. Its upper bound is close to, but below, both modern optima.

The label's own HVOD threshold converts the other way and lands almost exactly on FORUM's ceiling: 1,500 uMmin per dose x 16 = **98.5 mgh/L**, against FORUM's overexposure boundary of 98.0. A regulatory toxicity limit set decades ago and a 2026 trial-derived efficacy ceiling agree to within 0.5%, which is reassuring about the ceiling and says nothing about the floor.

The generalisable point: two published thresholds for the same drug, in incommensurable units, one per-dose and one cumulative, requiring a molecular weight and a dose count to compare. Anyone reading only one of them gets a self-consistent and incomplete answer.

What busulfan does to a baby that it does not do to an older child
  • Hepatic sinusoidal obstruction syndrome. Age under 1 year carried an odds ratio of 7.25 for SOS in a 105-patient paediatric HSCT cohort (univariate, p=0.037); it did not survive that study's multivariable model, where gemtuzumab exposure, ferritin >2400 ng/mL and INR >=1.3 did 5. Recorded here as a signal, not a confirmed independent risk factor. See risk-sinusoidal-obstruction-syndrome-in-infant-conditioning.
  • Incomplete ablation at the same exposure. In 105 children transplanted for non-malignant disease with myeloablative busulfan, age under 3 years was associated with mixed donor myeloid chimerism in 45.2% versus 1.9% at age >=3 years, with no difference in busulfan exposure between the groups 6. The same AUC does not do the same thing in a younger marrow. See risk-mixed-donor-chimerism-in-children-under-three.
  • Seizures. Anticonvulsant prophylaxis is mandatory from 12 h before the first dose to 24 h after the last, on the label and in Interfant-06 14.
  • Cardiac tamponade has been reported in paediatric patients with thalassaemia receiving high-dose oral busulfan with cyclophosphamide, preceded by abdominal pain and vomiting 4. The Interfant-06 regimen is high-dose oral busulfan with cyclophosphamide.
  • The label's dosing statement starts at 12 kg. The US intravenous label gives its dosing instruction 'for patients weighing more than 12 kg' 4. A six-month-old with KMT2A-rearranged ALL is typically 6-8 kg and therefore outside the sentence.
What this entry does not claim

No busulfan exposure-outcome analysis restricted to infants with KMT2A-rearranged ALL exists in the literature reached by this beat. FORUM's PK cohort starts at 0.5 years and is not stratified by fusion; Bartelink's cohort has a median age of 4.5 years; the chimerism and SOS cohorts are non-malignant and mixed-diagnosis respectively. Every infant-specific number above is an age finding borrowed across diagnoses, and the KMT2A-specific question is open.

A bounded absence, measured 2026-09-20 on PubMed: busulfan AND cyclophosphamide AND melphalan AND infant AND acute lymphoblastic leukemia AND conditioning AND Interfant returns 0 records. The identical query shape without the token Interfant returns 20, so the instrument composes and can return a one. The Bu/Cy/Mel conditioning regimen that Interfant-06 specifies has never been reported as a named analysis of its own.

Facts and figures
targetDNA - bifunctional alkylation at N7 of guanine producing intrastrand and interstrand cross-links, principally in haematopoietic stem and progenitor cells. No DNA-repair pathway node exists in this graph as a drug target.
modalitysmall molecule, bifunctional alkylsulfonate; NOT a prodrug - busulfan is directly reactive and is cleared by glutathione conjugation via GSTA1
role in this diseasemyeloablative backbone of irradiation-free conditioning for high-risk KMT2A-rearranged infant ALL proceeding to allogeneic HSCT in first remission
interfant06 conditioningoral busulfan 1.25 mg/kg every 6 h x 16 doses (5 mg/kg/day, days -7 to -4; 20 mg/kg total), then cyclophosphamide 60 mg/kg i.v. on days -3 and -2, then melphalan 140 mg/m2 i.v. on day -1; ATG-Fresenius 20 mg/kg on days -3 to -1 for matched non-sibling donors
label hvod thresholdAUC above 1,500 uM*min per dose = 98.5 mg*h/L cumulative over 16 doses
unit trapBusulfan exposure is published in two incommensurable units: uM*min PER DOSE on regulatory labels and older literature, mg*h/L CUMULATIVE in the modern trial literature. Conversion requires MW 246.3 and the number of doses. 1 uM*min = 0.004105 mg*h/L.
vocabularies searchedPubMed [tiab] counts on 2026-09-20: busulfan OR busulphan 6,367; Myleran 348; Busulfex 24; Busilvex 11. The generic names dominate retrieval and the trade names add almost nothing, so a generic-name search does NOT understate this drug.
optimal cumulative auc mg h per L
{
  "forum_2026": "73.3-98.0",
  "bartelink_2016": "78-101",
  "us_label_converted": "59.1-88.7 (from 900-1350 uM*min per dose x 16 doses, MW 246.3)"
}
infant specific
[
  "The US intravenous label's dosing instruction is written for patients weighing more than 12 kg; a 6-month-old with KMT2A-r ALL is typically 6-8 kg",
  "Label paediatric PK study used 1.0 mg/kg initial dose at age 4 years or under versus 0.8 mg/kg above 4 years",
  "Age under 1 year: univariate OR 7.25 for sinusoidal obstruction syndrome (not significant on multivariable adjustment)",
  "Age under 3 years: 45.2% versus 1.9% mixed donor myeloid chimerism at the same busulfan exposure, in non-malignant disease",
  "Interfant-06 route is ORAL, in a population where enteral absorption is variable and vomiting is common; the protocol permits intravenous substitution without specifying a dose conversion"
]
key ae
[
  "prolonged myelosuppression (boxed warning; 100% of patients)",
  "hepatic sinusoidal obstruction syndrome / veno-occlusive disease",
  "seizures - anticonvulsant prophylaxis mandatory",
  "mucositis",
  "gonadal failure and infertility",
  "pulmonary fibrosis ('busulfan lung')",
  "cardiac tamponade reported in paediatric thalassaemia on high-dose oral busulfan plus cyclophosphamide",
  "second malignancy (alkylator class)"
]
resistance
[
  "GSTA1 polymorphism alters clearance and therefore exposure - a pharmacokinetic rather than a tumour-cell resistance mechanism",
  "no KMT2A-rearrangement-specific busulfan resistance mechanism was found by this beat"
]
trials
[
  "NCT01949129"
]
How this could bite us back
  • Underexposure is a relapse mechanism that looks like a dosing detail. FORUM puts the relapse hazard ratio at 1.93 below 73.3 mg*h/L, and Interfant-06 makes the measurement that would detect underexposure OPTIONAL. An infant can be given a full protocol-compliant conditioning regimen and still be underconditioned, with no record that it happened.
  • The dose is per kilogram and the biology is not. An infant's busulfan clearance, marrow reserve and hepatic conjugating capacity are all changing month by month during the first year, so a mg/kg dose is a moving target in a way it is not in a ten-year-old.
  • Essential-function toxicity: busulfan ablates normal haematopoietic stem cells by design. Every long-term consequence of that - infertility, growth failure, second malignancy - is the intended mechanism operating on the wrong cells, and there is no selectivity to improve.
  • The mitigation for the main organ toxicity does not work. Prophylactic defibrotide failed its randomised trial for futility, so identifying an infant as high-risk for sinusoidal obstruction syndrome currently changes monitoring, not treatment.
  • Oral administration in a vomiting infant is an uncontrolled dose. The protocol's own hedge - measure levels 'if possible' - concedes that the delivered dose may not be the prescribed one, and in a drug with this therapeutic index that is the whole problem.
  • Busulfan conditioning may select for the residual clone rather than eliminate it: alkylator exposure is itself mutagenic, and the post-transplant relapses in this disease arise from cells that survived a myeloablative alkylator, which is a harsher selection than any targeted agent applies.
Still open
  1. Does the FORUM exposure window (73.3-98.0 mg*h/L) hold in infants under 12 months? The FORUM PK cohort starts at 0.5 years and is not age-stratified in its published analysis, so the window that now guides infant dosing has never been validated in infants.
  2. Interfant-06 specifies ORAL busulfan and permits intravenous substitution without a stated conversion. How many high-risk infants were actually conditioned orally, and did route affect exposure attainment? This is answerable from the Interfant-06 transplant dataset and has not been reported.
  3. Does the age-under-3 mixed-chimerism effect seen in non-malignant disease operate in infant KMT2A-rearranged ALL, where mixed chimerism would be a relapse risk rather than a partial cure?
  4. Is there a KMT2A-rearrangement-specific alkylator sensitivity? The one within-11q23 oxazaphosphorine difference recorded in this graph is an ifosfamide measurement, and no busulfan equivalent exists.
Every source behind this entry (6)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.book

    Interfant-06 specifies irradiation-free conditioning of oral busulfan 20 mg/kg total (1.25 mg/kg every 6 hours for four days), cyclophosphamide 60 mg/kg on each of two days and melphalan 140 mg/m2 on day -1, states that total body irradiation is not used because of its neurotoxic side effects in young children, requires seizure prophylaxis during busulfan, and makes busulfan blood-level monitoring conditional ('should be measured if possible').

    INTERFANT-06: International collaborative treatment protocol for infants under one year with acute lymphoblastic or biphenotypic leukemia (protocol document, SKION version 01, April 2006) · Pieters R, de Lorenzo P, Valsecchi MG, Peters C, and the Interfant-06 study group · Interfant-06 study group / SKION protocol document 2006
    What this evidence cannot show

    The full protocol PDF was retrieved and read on 2026-09-20; chapter 6.5 and section 1.3.9 are quoted directly. This is the protocol document, not a peer-reviewed report, and it is the 2006 SKION version - later amendments are not captured by it. It states an intention; it does not report what centres actually did, and in particular it does not tell us how many centres exercised the optional busulfan monitoring or used intravenous rather than oral busulfan.

    Evidence that argues against this (1)
    • The other large infant protocol chose a different irradiation-free regimen, and targeted busulfan rather than leaving monitoring optional. JPLSG MLL-10 used 'individualized targeted doses of busulfan, etoposide, and cyclophosphamide' in 38 of 43 protocol-specified transplants, achieving 3-year EFS 56.8% (95% CI 42.4-68.8) and OS 80.2% (95% CI 67.1-88.5) in high-risk KMT2A-rearranged infants. Two contemporaneous infant protocols therefore disagree on both the third agent (melphalan versus etoposide) and on whether busulfan exposure is targeted or merely watched. Takachi T et al. 2021
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In 145 paediatric ALL patients aged 0.5-19.5 years receiving fludarabine-busulfan-thiotepa conditioning within the randomised phase 3 FORUM trial, the optimal cumulative busulfan AUC was 73.3-98.0 mg*h/L; non-optimal exposure gave lower event-free survival (HR 1.93, p=0.008) and GVHD-free relapse-free survival (HR 2.01, p=0.002), with underexposure driving relapse (HR 1.93, p=0.019) and overexposure driving toxicity and GVHD.

    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 2026
    PMID 41779961DOI
    What this evidence cannot show

    A post hoc pharmacokinetic substudy of 145 of the FORUM participants who had busulfan levels available, not a randomised comparison of exposure targets; exposure was estimated from a population PK model rather than measured directly at every dose. The window was derived and evaluated in the same dataset. The lower age bound is 0.5 years, so infants are present but not separately analysed, and the cohort is paediatric ALL of all genotypes - no KMT2A-rearranged subgroup is reported.

    Evidence that argues against this (1)
    • A larger, older and independent cohort puts the optimum 5 mg*h/L higher at both ends: 78-101 mg*h/L in 674 children and young adults from 15 centres, with 2-year EFS 77.0% in the optimum band versus 66.1% at the low historical target of 58-86 mg*h/L. The two windows overlap but do not coincide, and 73.3-78 mg*h/L is 'optimal' on one analysis and 'low' on the other. A centre must choose which window to target and the choice is not settled by data. Bartelink IH et al. 2016
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In 674 children and young adults from 15 centres receiving busulfan-based conditioning (median age 4.5 years, IQR 1.4-10.7), 2-year event-free survival was 77.0% at a cumulative busulfan AUC of 78-101 mg*h/L versus 66.1% at the low historical target of 58-86 mg*h/L and 49.5% above 101 mg*h/L; transplantation-related mortality was 2.99-fold higher in the high-AUC group.

    Association of busulfan exposure with survival and toxicity after haemopoietic cell transplantation in children and young adults: a multicentre, retrospective cohort analysis. · Bartelink IH, Lalmohamed A, van Reij EM, Dvorak CC, Savic RM, Zwaveling J, Bredius RG, Egberts AC, et al. · The Lancet Haematology 2016
    PMID 27746112DOI
    What this evidence cannot show

    Retrospective and propensity-score-adjusted rather than randomised; 59% of the cohort had non-malignant disease, so the relapse endpoint is carried by a minority and the AUC-relapse relationship is estimated on 274 malignant cases. Exposure was recomputed centrally by NONMEM because the centres' own AUC estimates correlated poorly with the standardised method (r2=0.35), which is itself a warning that historical centre-reported AUCs are not comparable across studies.

    Evidence that argues against this (1)
    • The FORUM substudy, in a purely malignant and purely ALL population, shifts the window down to 73.3-98.0 mg*h/L and reports that patients inside it were statistically indistinguishable from matched TBI recipients in EFS, GRFS, OS and relapse. The disagreement is not trivial: 73-78 mg*h/L is the optimum floor on one analysis and underexposed on the other. Ben Hassine K et al. 2026
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    The US label for intravenous busulfan gives its dosing instruction for patients weighing more than 12 kg, reports a paediatric pharmacokinetic study (n=24, age 5 months to 16 years) targeting an AUC of 900-1350 uM*min with an initial dose of 1.0 mg/kg at age 4 years or under versus 0.8 mg/kg above 4 years, mandates anticonvulsant prophylaxis, warns of hepatic veno-occlusive disease above an AUC of 1,500 uM*min, and notes cardiac tamponade in paediatric thalassaemia patients receiving high-dose oral busulfan with cyclophosphamide.

    BUSULFAN injection, for intravenous use - US prescribing information (Structured Product Label, ANDA215102) · Eugia US LLC · DailyMed / FDA Structured Product Labeling, openFDA drug label API 2024
    What this evidence cannot show

    Retrieved from the openFDA label API on 2026-09-20; SPL set id 06883ac0-4b2a-4f26-89bf-512d37b9ba7f, effective 2024-06-27. This is a generic intravenous product's label and inherits its paediatric section from the reference listed drug; it is a regulatory document, so it states what is approved rather than current best practice, and the 24-patient PK study it describes is decades old. The unit conversion from uM*min to mg*h/L in this entry was performed by the beat using busulfan MW 246.3 and a 16-dose schedule; it is arithmetic on the label's stated numbers, not a figure the label itself prints.

    Evidence that argues against this (1)
    • The label's paediatric target is not wrong so much as superseded, and reading it as a defect would be unfair: at 900-1350 uM*min per dose over 16 doses it converts to 59-89 mg*h/L, whose UPPER portion overlaps both modern optima. The FORUM analysis also reports that patients inside the modern window were indistinguishable from matched TBI recipients, which is a claim about the window and not about the label. The honest statement is that a centre dosing to the bottom of the labelled range would be underexposing by current evidence, not that the label is unsafe. Ben Hassine K et al. 2026
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In 105 children, adolescents and young adults undergoing allogeneic HSCT at a single paediatric centre, 15 (14.3%) developed hepatic sinusoidal obstruction syndrome with 20.0% mortality; patient age under 1 year carried an odds ratio of 7.25 (p=0.037) in univariate analysis, but did not remain significant in the multivariable model.

    PMID 34255148DOI
    What this evidence cannot show

    Single centre, 105 patients over 12 years, 15 SOS events. The number of patients under 1 year is not stated in the abstract, so the odds ratio of 7.25 rests on a small numerator and its confidence interval is wide. It is a univariate finding that did NOT survive multivariable adjustment, which is the honest headline; the factors that did survive were gemtuzumab ozogamicin, ferritin above 2400 ng/mL and INR at or above 1.3. This entry records it as an age signal, not an established independent risk factor.

    Evidence that argues against this (1)
    • The intervention that would follow from an SOS risk signal has been tested and failed. In HARMONY, an open-label randomised phase 3 trial in 372 patients at high or very high risk of sinusoidal obstruction syndrome (median age 14.0 years, paediatric patients from 1 month of age), prophylactic defibrotide gave day-30 SOS-free survival of 67% versus 73% on best supportive care alone (HR 1.27, 95% CI 0.84-1.93; p=0.85), and enrolment was stopped early for presumed futility. Identifying infants as high-risk does not currently license giving them anything. Grupp SA et al. 2023
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In 105 children with non-malignant disease receiving myeloablative busulfan-based conditioning, the 3-year cumulative incidence of mixed donor myeloid chimerism was 45.2% in those aged under 3 years versus 1.9% in those aged 3 years or older (p<0.001), with no difference in busulfan exposure between the age groups.

    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 2025
    PMID 39938805DOI
    What this evidence cannot show

    Non-malignant disease only, retrospective, single-institution, 105 patients over 16 years, and the age cut-off of 2.9 years was derived from ROC analysis within the same dataset rather than prespecified. Mixed chimerism matters differently in leukaemia than in a genetic disease - in leukaemia it is a relapse risk and a trigger for immunological intervention, not a partial correction. Whether the same age effect operates in an infant transplanted for KMT2A-rearranged ALL is untested. The proposed mechanism, greater marrow reserve in younger children, is the authors' hypothesis and is not measured.

    Evidence that argues against this (1)
    • The same study reports no impact of age or busulfan exposure on immunologic graft rejection, acute or chronic GVHD, transplant-related mortality, or sinusoidal obstruction syndrome. So the age effect is specific to the degree of myeloid replacement and does not propagate to the outcomes a leukaemia transplant is judged on. Reading it as 'young children do worse after busulfan' would overstate it. Dvorak CC et al. 2025
Connected to