Weight or surface area - the Interfant-06 conditioning week uses both units in eight days, and the one measured correction in the literature went the wrong way
also called mg/kg versus mg/m2 · body surface area dosing · body weight dosing · BSA dosing infants · conditioning dose units · 30 mg/m2 equals 1 mg/kg rule
Chemotherapy doses are worked out either from a child's weight or from their body surface area, and the two give different answers in a baby because babies are a different shape from adults. The transplant conditioning week for an infant uses both systems in eight days, and the only time anyone measured whether a size correction actually worked, it turned out to have pushed the dose the wrong way.
The same thing, for a scientist
Body surface area and body weight diverge most in the smallest patients: surface area per kilogram is roughly twice as high in a neonate as in an adult, so a mg/m2 dose delivers substantially more drug per kilogram to an infant than to an older child, while a mg/kg dose delivers less per unit of surface area. Interfant-06's conditioning regimen uses both conventions inside one eight-day block: busulfan 5 mg/kg/day, cyclophosphamide 60 mg/kg, melphalan 140 mg/m2, ATG 20 mg/kg [E1]. The protocol's single stated size adjustment is for obesity, not for infancy - 'For very obese children, the conditioning should be calculated not according to the actual body weight, but to the 97th percentile of the current age' [E1]. The only published measurement of whether such a correction works found it overcorrected: in 85 paediatric transplants, obese patients dosed with melphalan on adjusted ideal body weight had significantly LOWER predicted exposure (p=0.02), and the same programme applies a separate unvalidated adjustment at age under 2 years [E2]. The modern answer is neither unit but exposure: model-based fludarabine dosing beat body-surface-area dosing on graft rejection, 1.8% versus 8.5% [E3].
One week, two units1 deeper
| day | agent | dose | unit |
|---|---|---|---|
| -7 to -4 | busulfan | 5 mg/kg/day (1.25 mg/kg q6h) | weight |
| -3, -2 | cyclophosphamide | 60 mg/kg | weight |
| -3 to -1 (MD only) | ATG-Fresenius | 20 mg/kg | weight |
| -1 | melphalan | 140 mg/m2 | surface area |
| +1, +3, +6 | methotrexate (GVHD prophylaxis) | 10 mg/m2 | surface area |
All from Interfant-06 chapter 6 1. Nothing in the protocol explains the mixture, and it is almost certainly inherited: each agent carries the convention of the trial in which its conditioning dose was first established.
The practical consequence in a 7 kg seven-month-old, whose body surface area is roughly 0.35 m2: the mg/kg agents are scaled to a number that grows roughly linearly with size, and melphalan is scaled to a number that does not. The same child is being sized by two incompatible rules in the same week, and no one reconciles them.
Why the two units diverge in an infant specifically
Surface area scales approximately with weight to the power of two-thirds, so the ratio of surface area to weight rises as size falls. A neonate carries roughly twice the surface area per kilogram of an adult. The familiar clinical shorthand - divide a mg/m2 dose by 30 to get mg/kg - is derived for a child of about 30 kg and is progressively wrong below that.
Which unit is correct depends on which physiological process governs the drug's disposition, and this differs by agent:
- Renally cleared drugs track glomerular filtration rate, which correlates better with surface area in older children but is still maturing in the first year, so neither unit is right below 12 months.
drug-fludarabineis the case in point. - Hepatically metabolised drugs track liver mass and enzyme expression; liver mass per kilogram is higher in infants while enzyme expression is lower and developmentally staged.
drug-busulfan, cleared by glutathione conjugation, sits here. - Drugs consumed by their target track the target's mass, which is neither.
drug-antithymocyte-globulinis consumed by the lymphocytes it destroys.
The one time anyone checked a size correction, it was wrong
Dvorak and colleagues applied a pharmacokinetic model to 85 paediatric transplants conditioned with melphalan. Most patients received 140 mg/m2 on actual body weight; 'others received a dose adjusted for obesity or age <2 years' 2.
For the obesity adjustment, the model 'predicted lower melphalan exposure (P = .02) in obese patients where adjusted ideal body weight was utilized, suggesting overcorrection of the dose' 2.
That correction was applied for exactly the reasons an infant correction is applied: a physiological argument that the standard size metric misrepresents the volume of distribution in an atypical body composition. It was reasonable, it was widely used, and when someone finally measured the exposure it produced, it was wrong in a statistically detectable direction.
The infant adjustment in the same programme has never been measured. It is currently in the position the obesity adjustment occupied before that analysis: physiologically motivated, clinically routine, and unvalidated.
The answer the field is converging on is neither unit
Both mg/kg and mg/m2 are proxies for exposure. Where exposure can be measured or modelled, the proxy becomes unnecessary - and where it has been compared, exposure wins:
| agent | proxy dosing | exposure-guided dosing | result |
|---|---|---|---|
| fludarabine | body surface area | model-based to cAUC 20 mg*h/L | graft rejection 8.5% -> 1.8%; multivariable HR for BSA dosing 5.12 3 |
| busulfan | mg/kg by weight band | TDM to cAUC 73.3-98.0 mg*h/L | non-optimal exposure: EFS HR 1.93, GRFS HR 2.01 4 |
| melphalan | 140 mg/m2 | not established | exposure above 6.81 mg*h/L associated with more acute GVHD 2 |
| treosulfan | 10 g/m2/day fixed | proposed 4,800 mg*h/L, not adopted | mortality HR 1.46 per 1,000 mg*h/L |
Every row where the comparison has been made favours exposure over the size proxy. And this is where Interfant-06's central weakness sits: it mandates therapeutic drug monitoring for ciclosporin, an immunosuppressant, and makes it conditional for busulfan, the agent with the narrow window and the veno-occlusive risk 1.
| only stated size adjustment | obesity: calculate to the 97th weight percentile for current age rather than actual body weight. NO age-based adjustment appears anywhere in the conditioning chapter. |
|---|---|
| the measured failure | In 85 paediatric transplants, obesity-adjusted melphalan dosing on adjusted ideal body weight produced significantly LOWER model-predicted exposure (p=0.02) - the correction overcorrected. The analogous age-under-2 adjustment in the same programme has never been measured. |
| the protocol irony | Interfant-06 mandates ciclosporin level monitoring with dose adjustment and makes busulfan level monitoring conditional on being 'possible' - therapeutic drug monitoring is required for the immunosuppressant and optional for the agent with the narrow window and the veno-occlusive risk. |
interfant06 conditioning units
{
"busulfan": "5 mg/kg/day (1.25 mg/kg q6h) - WEIGHT",
"cyclophosphamide": "60 mg/kg x2 - WEIGHT",
"atg_fresenius": "20 mg/kg x3 - WEIGHT",
"melphalan": "140 mg/m2 x1 - SURFACE AREA",
"methotrexate_gvhd_prophylaxis": "10 mg/m2 on days +1/+3/+6 - SURFACE AREA"
}scaling facts
[ "Body surface area scales approximately with weight^(2/3), so surface area per kilogram rises as size falls", "A neonate carries roughly twice the surface area per kilogram of an adult", "The clinical shorthand 'divide mg/m2 by 30 to get mg/kg' is derived for a ~30 kg child and is progressively wrong below that" ]
which unit is right depends on the agent
{
"renally_cleared": "fludarabine - tracks glomerular filtration rate, still maturing through the first 1-2 years, so neither unit is correct below 12 months",
"hepatically_cleared": "busulfan - glutathione conjugation; liver mass per kilogram is higher in infants while conjugating enzyme expression is lower and developmentally staged",
"target_mediated": "anti-thymocyte globulin - consumed by the lymphocytes it destroys, so it tracks lymphocyte mass, which is neither weight nor surface area"
}exposure beats proxy where compared
{
"fludarabine": "BSA 8.5% rejection vs model-based 1.8%; multivariable HR 5.12 for BSA dosing",
"busulfan": "non-optimal exposure EFS HR 1.93, GRFS HR 2.01",
"melphalan": "exposure >6.81 mg*h/L associated with more acute GVHD; no established target",
"treosulfan": "mortality HR 1.46 per additional 1,000 mg*h/L; proposed target 4,800 mg*h/L not adopted"
}- A physiologically motivated correction can be worse than none. The one time a size adjustment in paediatric conditioning was measured against modelled exposure, it moved patients further from target, and it had been applied routinely for years on reasoning that sounded right.
- The mixture of units hides the problem. Because each agent carries the convention of its own registration trial, nobody sees the regimen as a whole, and no single document asks what dose of each drug a 7 kg patient should receive.
- Exposure targeting relocates the arbitrariness rather than removing it. Two published busulfan windows differ by about 5 mg*h/L at both ends, so a centre adopting therapeutic drug monitoring still has to choose which paper to believe.
- Monitoring infrastructure is a determinant of outcome, and it is unevenly distributed. 'Measured if possible' means the infants in well-resourced centres are receiving a different intervention from those elsewhere on the same protocol, and the trial cannot detect the difference.
- The protocol adjusts frontline chemotherapy for age and does not adjust conditioning for age. If the reasoning that produced two-thirds dosing for cyclophosphamide under 6 months is right, it does not stop being right on day -3 of a transplant.
- What is the age-under-2 melphalan dose adjustment, in which direction, and does it produce the intended exposure? The only comparable correction in the same programme was measured and found to overcorrect, and this one has not been measured at all.
- Is there any unit that is correct for an infant, or is the honest answer that every conditioning agent needs exposure-guided dosing below 12 months? Renal, hepatic and target-mediated clearance are all maturing simultaneously and none of them tracks either weight or surface area.
- Would mandating busulfan therapeutic drug monitoring, rather than permitting it, change infant transplant outcomes? Interfant-06 made it optional; JPLSG MLL-10 targeted it; the two protocols differ on many axes and this one has never been isolated.
- How many high-risk Interfant-06 infants actually had a busulfan level measured? If the answer is recoverable from trial records it would convert the largest infant conditioning dataset into an exposure-response analysis at no additional cost.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.book
The Interfant-06 conditioning block doses busulfan at 5 mg/kg/day, cyclophosphamide at 60 mg/kg, ATG-Fresenius at 20 mg/kg and melphalan at 140 mg/m2, with post-transplant methotrexate at 10 mg/m2; its only stated size adjustment is for obesity ('For very obese children, the conditioning should be calculated not according to the actual body weight, but to the 97th percentile of the current age'); ciclosporin levels are mandated to be measured and the dose adjusted, while busulfan levels '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 2006What this evidence cannot show
Chapters 6.5 and 6.6 of the 2006 SKION document, read on 2026-09-20. The frontline chapters of the same protocol DO contain infant dose reductions by age for chemotherapy (two-thirds under 6 months, three-quarters from 6 to 12 months for cyclophosphamide, per this graph's existing backbone entries), so the absence of such a reduction in the conditioning chapter is a difference between chapters and may reflect that conditioning is written by a different committee rather than a considered decision. A protocol states intent, not practice, and what individual transplant centres actually did is unrecorded.
Evidence that argues against this (1)
- Mixing units within a regimen is standard across oncology and is not by itself an error; each agent's dose was established in the unit its own registration or conditioning trial used, and changing it would invalidate the dose. The criticism that bites is not the mixture but the absence of any age-based adjustment in the conditioning chapter of a protocol whose entire population is under one year old, while the same protocol adjusts frontline chemotherapy by age. Dvorak CC et al. 2025
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In 85 paediatric melphalan-conditioned transplants, 61.2% received 140 mg/m2 on actual body weight while others received a dose adjusted for obesity or for age under 2 years; a pharmacokinetic model predicted significantly LOWER melphalan exposure in obese patients dosed on adjusted ideal body weight (p=0.02), indicating overcorrection, and exposure above the cohort median of 6.81 mg*h/L was associated with increased acute GVHD.
High Melphalan Exposure Increases the Risk of Graft-Versus-Host Disease in Pediatric Patients Undergoing Alpha-Beta T-Cell Depleted Haploidentical Transplantation. · Dvorak CC, Cho S, Salinas Cisneros G, Higham CS, Chu J, Winestone LE, Temple WC, Kharbanda S, et al. · Transplantation and Cellular Therapy 2025What this evidence cannot show
Exposure was MODEL-PREDICTED rather than measured, so 'overcorrection' is a statement about a model's output in obese patients and depends on the model being right in exactly the body composition where models are least reliable. 85 patients, retrospective, alpha/beta T-cell-depleted haploidentical setting. The age-under-2 adjustment is mentioned but its magnitude, direction and the number of patients receiving it are not given, so this entry can say only that such an adjustment exists and is unvalidated.
Evidence that argues against this (1)
- A model that finds a correction overcorrects may itself be miscalibrated in the corrected population - obese body composition is precisely where allometric scaling assumptions break down, so a model built mainly on non-obese patients predicting low exposure in obese ones is at least as consistent with model misspecification as with genuine overcorrection. The finding argues for measuring melphalan exposure directly, not for abandoning the adjustment. Dvorak CC et al. 2025
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 410 paediatric and young-adult allogeneic transplants, fludarabine dosed by pharmacokinetic model to a cumulative AUC of 20 mg*h/L gave immunologic graft rejection in 1.8% versus 8.5% with body-surface-area dosing (p=0.004), with a multivariable hazard ratio for BSA dosing of 5.12 (95% CI 1.51-17.28, p=0.009).
Fludarabine Pharmacokinetic Model-Based Dosing Minimizes the Risk of Graft Rejection in Pediatric Patients Undergoing Allogeneic Hematopoietic Cell Transplantation. · Dvorak CC, Waters E, Temple WC, Salinas Cisneros G, Chu J, Winestone LE, Higham CS, Kharbanda S, et al. · Transplantation and Cellular Therapy 2026What this evidence cannot show
Retrospective single-institution comparison across 2008-2024, with model-based dosing adopted later than BSA dosing, so the comparison is confounded with calendar time and every concurrent improvement in transplant practice. Twenty-four rejection events support a five-variable multivariable model and the confidence interval is correspondingly wide. Malignant and non-malignant indications pooled; the authors report the benefit was concentrated in patients carrying two or three other rejection risk factors.
Evidence that argues against this (1)
- Exposure-guided dosing is not free and is not universally available. It requires assay access, a validated population model, and a turnaround fast enough to adjust a dose mid-conditioning - infrastructure that Interfant-06 implicitly acknowledges is unevenly distributed when it writes 'blood levels of the drug should be measured if possible'. A demonstration that exposure dosing is better in a well-resourced single centre does not establish that mandating it internationally would improve outcomes. Pieters R et al. 2006
- E4strong 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 within the FORUM trial, busulfan exposure outside the optimal cumulative AUC window of 73.3-98.0 mg*h/L was associated with lower event-free survival (HR 1.93, p=0.008) and lower GVHD-free relapse-free survival (HR 2.01, p=0.002).
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
Post hoc pharmacokinetic substudy of 145 of the FORUM participants with available busulfan levels, using model-estimated rather than directly integrated exposure, with the window derived and evaluated in the same dataset. Not a randomised comparison of dosing strategies. Age range starts at 0.5 years but no infant-stratified result is reported, and no KMT2A-rearranged subgroup is given.
Evidence that argues against this (1)
- An earlier and larger cohort of 674 children puts the optimal window at 78-101 mg*h/L rather than 73.3-98.0. Exposure-guided dosing therefore requires choosing which published window to target, and the two disagree by about 5 mg*h/L at both ends - so replacing a size proxy with an exposure target does not eliminate the arbitrariness, it relocates it. Bartelink IH et al. 2016
- part of →Interfant-06 — 651 infants, and the definitive answer that myeloid-style consolidation does not help
the conditioning block mixes mg/kg and mg/m2 in eight days and carries no age adjustment
the only BSA-dosed cytotoxic in the regimen, and the one whose obesity correction was measured and found to overcorrect
model-based dosing beat BSA dosing on graft rejection, 1.8% vs 8.5%
the agent with the narrow window whose monitoring Interfant-06 made optional while mandating ciclosporin levels
which size unit is right depends on which maturing organ system clears the agent