Fludarabine - the immunosuppressive half of modern conditioning, given as a phosphate prodrug, cleared by the kidney, and dosed by a body-surface-area rule that measurably fails in children
also called fludarabine · fludarabin · FLU · Flu · fludarabine phosphate (the administered species) · fludarabine monophosphate · 2-fluoro-ara-AMP · F-ara-AMP · F-ara-A (the circulating dephosphorylated species) · 9-beta-D-arabinofuranosyl-2-fluoroadenine · F-ara-ATP (the intracellular active species) · Fludara (trade name) · purine nucleoside analogue (class)
Fludarabine is not there to kill the leukaemia; it is there to suppress the child's own immune system so the donor's cells are not rejected. The dose is worked out from body size, and in children that rule is measurably wrong often enough that dosing by measured blood levels instead cuts the graft-rejection rate from 8.5% to 1.8%.
The same thing, for a scientist
Fludarabine is a purine nucleoside analogue administered as fludarabine monophosphate, a prodrug that is dephosphorylated in plasma to F-ara-A and re-phosphorylated intracellularly by deoxycytidine kinase to the active F-ara-ATP [E1]. Its role in conditioning is lymphodepletion rather than antileukaemic cytotoxicity: it is profoundly T-cell suppressive with minimal NK activity, and it exists in the regimen to prevent rejection of the donor graft [E2]. It is the fludarabine in BU/THIO/FLU and TREO/THIO/FLU, the two chemoconditioning backbones compared in FORUM, and is the mandated partner in the US treosulfan label. In 410 paediatric transplants, pharmacokinetic model-based dosing to a cumulative AUC of 20 mg*h/L reduced immunologic rejection from 8.5% to 1.8% against body-surface-area dosing (multivariable HR for BSA dosing 5.12, 95% CI 1.51-17.28, p=0.009) [E2]. Because F-ara-A is renally cleared, its exposure in an infant is governed by a glomerular filtration rate that is still maturing and that creatinine-based estimates measure badly - a physiologically based model of paediatric fludarabine predicted exposure better when informed by NUCLEAR GFR than by estimated GFR [E1].
Three molecules, one name
This is the prodrug case the graph's backbone beat flagged, in its cleanest form:
| species | where | what it is |
|---|---|---|
| fludarabine monophosphate (F-ara-AMP, 2-fluoro-ara-AMP) | the vial and the infusion | the administered prodrug; water-soluble because of the phosphate |
| F-ara-A | plasma | dephosphorylated by plasma 5'-nucleotidase; this is the species measured by therapeutic drug monitoring and the one that is renally cleared |
| F-ara-ATP | inside the lymphocyte | re-phosphorylated by deoxycytidine kinase; incorporates into DNA, inhibits DNA polymerase, primase, ribonucleotide reductase and DNA ligase I |
All three are called 'fludarabine' in clinical prose. A paper reporting a 'fludarabine AUC' means F-ara-A in plasma, not the infused prodrug and not the active triphosphate, and the target of 20 mg*h/L 2 is a concentration of a molecule that is neither given nor active. mechanism-fludarabine-prodrug-chain holds the activation chain in full.
The practical consequence is that the drug's variability has two separate sources that do not move together: renal clearance determines F-ara-A exposure, and deoxycytidine kinase activity determines how much of that exposure becomes active drug inside a cell. Monitoring addresses only the first.
What fludarabine is actually for, and why that changes how a failure looks1 deeper
Fludarabine's contribution to conditioning is 'profound immunosuppressive effects on T cells, albeit with minimal activity against NK cells', and it is in the regimen 'to prevent rejection of donor cells' 2. It is not carrying the antileukaemic load - the alkylator is.
So underexposure to fludarabine does not present as relapse. It presents as graft rejection, days to weeks after transplant, in a child who has already had their marrow ablated. In the Dvorak cohort, five factors independently predicted rejection: a pre-lymphodepletion absolute lymphocyte count at or above 1.275 x10^9/L (9.4% vs 1.1%), a female donor for a male recipient (10.9% vs 4%), an HLA-mismatched donor (9.7% vs 1.6%), CMV seropositivity in donor or recipient (7.3% vs 1.6%), and body-surface-area rather than model-based fludarabine dosing (8.5% vs 1.8%) 2.
Of those five, four are fixed at the time the transplant is planned. The fifth is a dosing decision.
Why body-surface-area dosing fails hardest in the smallest patients
F-ara-A is cleared renally, so its exposure tracks glomerular filtration rate. Body surface area is a geometric proxy for GFR, and the proxy degrades where the ratio of surface area to renal function is least typical - which is infancy, when GFR is still maturing towards adult values over the first one to two years.
A physiologically based model of paediatric fludarabine makes the point sharply: prediction of individual exposure improved when the model was informed by nuclear GFR rather than estimated GFR 1. Estimated GFR in an infant is computed from serum creatinine, and creatinine is generated by muscle. An infant has very little muscle, so the estimate that stands in for the variable that determines the dose is itself least reliable in the age group where the dose matters most. See concept-infant-organ-immaturity-and-conditioning-drug-clearance.
Where fludarabine is and is not in this disease
It is not in Interfant-06. The 2006 protocol's conditioning is busulfan, cyclophosphamide and melphalan, with ATG for non-sibling donors, and fludarabine appears nowhere in the conditioning chapter 3. Nor is it in JPLSG MLL-10's busulfan/etoposide/cyclophosphamide.
It is in everything since. BU/THIO/FLU and TREO/THIO/FLU are the two chemoconditioning regimens FORUM compared, and the US treosulfan approval is specifically in combination with fludarabine 4. The substitution that produced them is described plainly in the standing review: 'Busulfan with cyclophosphamide is a long-established alternative to TBI-based conditioning in ALL patients. Substituting fludarabine for cyclophosphamide reduces toxicity, but may not be as effective, prompting the addition of a third agent, such as thiotepa, melphalan, and now clofarabine' 5.
Read that sequence carefully. Fludarabine entered conditioning as a toxicity-reduction substitution for cyclophosphamide, was suspected of costing efficacy, and a third agent was added to make up the difference. The modern three-drug regimen is the residue of that trade, not a design from first principles - and none of the three steps was tested in an infant.
| target | DNA synthesis in lymphocytes - F-ara-ATP is incorporated into DNA and inhibits DNA polymerase alpha, primase, ribonucleotide reductase and DNA ligase I. The function delivered in conditioning is host T-cell depletion to prevent graft rejection, not antileukaemic cytotoxicity. |
|---|---|
| modality | small molecule purine nucleoside analogue, administered as a PHOSPHATE PRODRUG (fludarabine monophosphate) |
| activation chain | fludarabine monophosphate (given) -> plasma 5'-nucleotidase -> F-ara-A (circulating, renally cleared, the species measured by TDM) -> deoxycytidine kinase -> F-ara-ATP (intracellular, active) |
| role in this disease | lymphodepletion in modern chemoconditioning (BU/THIO/FLU and TREO/THIO/FLU); NOT a component of Interfant-06 or JPLSG MLL-10 infant conditioning |
| target cumulative auc mg h per L | 20 |
| what the auc measures | F-ara-A in plasma - neither the administered prodrug nor the active triphosphate |
| elimination | renal; exposure is governed by glomerular filtration rate, which is still maturing through the first 1-2 years of life |
| bsa vs model based dosing | immunologic graft rejection 8.5% with body-surface-area dosing versus 1.8% with model-based dosing in 410 paediatric transplants; multivariable HR 5.12 (1.51-17.28) |
| vocabularies searched | PubMed [tiab] on 2026-09-20: fludarabine OR fludarabin OR 'F-ara-A' OR '2-fluoro-ara-AMP' 6,828; Fludara 40. Generic-name retrieval is adequate. |
infant specific
[ "Renal clearance in the first year is a moving target, so a fixed mg/m2 dose produces a changing exposure across the age band Interfant-06 treats as one category", "Estimated GFR is creatinine-based and creatinine derives from muscle mass, so the covariate that governs the dose is estimated worst in the patient with the least muscle; a PBPK model predicted exposure better from NUCLEAR GFR than estimated GFR", "Underexposure presents as GRAFT REJECTION in an already-ablated infant, not as relapse - a failure mode with a very short fuse" ]
key ae
[ "profound and prolonged T-cell lymphopenia with late opportunistic infection", "neurotoxicity at high cumulative exposure, including a rare delayed leucoencephalopathy", "myelosuppression", "autoimmune haemolytic anaemia", "transfusion-associated GVHD risk requiring irradiated blood products" ]
resistance
[ "low deoxycytidine kinase activity reduces conversion of F-ara-A to the active F-ara-ATP - a resistance mechanism invisible to plasma exposure monitoring", "5'-nucleotidase NT5C2 dephosphorylation, which is also a recognised relapse-associated mechanism in ALL for thiopurines", "no KMT2A-rearrangement-specific fludarabine measurement was found by this beat" ]
trials
[ "NCT01949129" ]
- The agent that protects the graft also disarms it. Fludarabine is in the regimen to stop the host rejecting the donor, and it does that by destroying T cells - the same effector class that any graft-versus-leukaemia effect must run through. The trade-off is structural and is never quantified in an infant.
- Underexposure kills fast and silently. A fludarabine shortfall shows up as graft rejection in a child whose own marrow has already been ablated, which is a medical emergency with no natural recovery path, unlike a relapse that presents over weeks.
- The measured species is the wrong species. Therapeutic drug monitoring reports F-ara-A in plasma; the drug that works is F-ara-ATP inside a lymphocyte, and nothing in routine practice measures the conversion step or the kinase that performs it.
- The covariate that determines the dose is estimated worst in the patient who needs it most. eGFR depends on creatinine, creatinine depends on muscle, and an infant has almost none - so the standard renal adjustment in this age group is close to uninformative.
- Prolonged T-cell lymphopenia in a baby sits on top of B-cell aplasia from any prior CD19-directed therapy and on top of GVHD immunosuppression, producing a compound immunodeficiency in a patient with no pre-existing immunological memory to fall back on.
- What fludarabine exposure does a 7 kg six-month-old actually achieve on 30 mg/m2/day? No published paediatric fludarabine exposure analysis separates infants, and renal maturation makes this the age band where a body-surface-area rule is least defensible.
- Monitoring measures F-ara-A but the active species is F-ara-ATP, generated by deoxycytidine kinase inside the cell. Does deoxycytidine kinase activity vary enough between infants to make plasma exposure a poor proxy for lymphodepletion?
- Would replacing cyclophosphamide with fludarabine in an INFANT conditioning regimen reduce toxicity without the relapse cost that FORUM found in older children? The substitution has never been tested below 4 years of age.
- KMT2A-rearranged infant blasts are resistant to TRAIL and relatively insensitive to FasL. Does profound fludarabine-induced T-cell depletion additionally blunt whatever graft-versus-leukaemia effect exists in this disease, making the immunosuppressive agent a contributor to relapse?
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Fludarabine is administered as the prodrug fludarabine monophosphate and shows substantial pharmacokinetic variability; in a physiologically based pharmacokinetic virtual-twin analysis of 28 paediatric HSCT patients, informing the model with plasma protein scaling factor and NUCLEAR glomerular filtration rate predicted individual exposure better than the model using estimated GFR (mean fold error 0.91 total, 0.88 unbound).
Physiologically Based Pharmacokinetic Virtual Twin Approach for Fludarabine Dosing in Pediatric Hematopoietic Stem Cell Transplantation. · Mseddi M, Nath C, Ben Hassine K, Mlakar V, Gloor Y, Gonzales F, Uppugunduri CRS, Daali Y, et al. · CPT: Pharmacometrics & Systems Pharmacology 2026What this evidence cannot show
A modelling study validated in only 28 paediatric patients, with no clinical outcome endpoint - better exposure prediction is not demonstrated to produce better transplants. The model was built 'middle-out' from literature data, so its structural assumptions are inherited rather than fitted. The age distribution of the 28 patients is not given in the abstract and infants are not separately reported, so the inference that the eGFR problem is worst in infancy is this beat's physiological reasoning and not a finding of the paper. Nuclear GFR is not routinely available before transplant in most centres, which limits the practical reach of the result.
Evidence that argues against this (1)
- Better individual exposure prediction has value only if exposure is what matters, and the outcome evidence for fludarabine exposure targeting is retrospective. The largest clinical dataset compares model-based dosing against body-surface-area dosing rather than testing exposure as a continuous predictor, and its benefit was concentrated in patients who already carried two or three other rejection risk factors - so a universal precision-dosing claim outruns the data. Dvorak CC et al. 2026
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In 410 paediatric and young-adult first allogeneic transplants (median age 8 years, range 0.1-26.5) with fludarabine-containing conditioning, model-based dosing to a cumulative AUC of 20 mg*h/L was associated with immunologic graft rejection in 1.8% versus 8.5% with body-surface-area dosing (p=0.004; multivariable hazard ratio for BSA dosing 5.12, 95% CI 1.51-17.28, p=0.009), alongside four other independent rejection risk factors.
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 a 16-year period (2008-2024) in which model-based dosing was adopted later than BSA dosing, so the comparison is confounded with calendar time and with every other improvement in supportive care, donor selection and GVHD prophylaxis over that interval. The overall rejection rate is 5.8% (24 events in 410 patients), so the multivariable model estimates five coefficients on roughly two dozen events and the confidence interval for the BSA hazard ratio is correspondingly wide (1.51-17.28). Both malignant and non-malignant indications are pooled. The authors themselves note the benefit was most apparent in patients with two or three other risk factors.
Evidence that argues against this (1)
- The authors' own framing limits the claim: the benefit of model-based dosing 'was most significant in patients with other risk factors for rejection', and they justify universal adoption on the SIMPLICITY of the approach rather than on demonstrated universal benefit. In a low-risk recipient of a 10/10 matched graft, the absolute rejection rate on BSA dosing is already 1.6%, and there is no room in that number for a five-fold hazard reduction to be clinically meaningful. Dvorak CC et al. 2026
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.book
The Interfant-06 conditioning regimen contains no fludarabine: it is oral busulfan days -7 to -4, cyclophosphamide days -3 and -2, melphalan day -1, with ATG-Fresenius 20 mg/kg on days -3 to -1 for matched non-sibling donors only.
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
Read directly from chapter 6.5 of the retrieved protocol PDF on 2026-09-20. This is the 2006 SKION version; later amendments were not retrieved and it is possible that conditioning was updated during the trial's long accrual. It also states intent, not practice - centres may have substituted fludarabine-containing regimens in individual cases, and no report of the Interfant-06 transplanted subgroup describes the conditioning actually delivered. An absence in a protocol document is therefore a bounded absence about the protocol, not about the patients.
Evidence that argues against this (1)
- The regimen that Interfant-06 omits has since become the international standard for paediatric chemoconditioning, and the substitution that created it was explicitly motivated by toxicity: 'Substituting fludarabine for cyclophosphamide reduces toxicity, but may not be as effective, prompting the addition of a third agent.' Interfant-06's cyclophosphamide-containing conditioning is therefore the older and more toxic arm of a comparison the field has since made, without any infant being enrolled in it. Ben Hassine K et al. 2021
- E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
The US treosulfan approval (GRAFAPEX, NDA 214759, 2025-01-21) is specifically for use in combination with fludarabine, with fludarabine 30 mg/m2/day intravenously on days -6 through -2 alongside treosulfan 10 g/m2/day on days -4, -3 and -2.
GRAFAPEX (treosulfan) for injection - US prescribing information, dosage regimen table (NDA 214759; SPL set id 2d06b03c-17aa-0492-e063-6294a90abd8f) · Medexus Pharma, Inc.; US Food and Drug Administration · openFDA Drug Label API, api.fda.gov 2025What this evidence cannot show
Retrieved live on 2026-09-20. The label's fludarabine dose is body-surface-area based and fixed, with no provision for exposure-guided adjustment, which is in direct tension with the paediatric evidence in E2 - regulatory text lags practice and should not be read as a recommendation against monitoring. The indication is AML and MDS in patients 1 year and older, so the regimen as labelled does not cover an infant with KMT2A-rearranged ALL.
Evidence that argues against this (1)
- A labelled BSA-based fludarabine dose is exactly the dosing strategy that performed worse in the largest paediatric analysis, where BSA dosing carried a multivariable rejection hazard ratio of 5.12 against model-based dosing. The label is a regulatory minimum and not a statement of best practice, and treating it as the latter is the scope error this project's brief warns about for statutes and database facts. Dvorak CC et al. 2026
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review
Fludarabine entered myeloablative conditioning as a toxicity-reducing substitute for cyclophosphamide alongside busulfan, at a suspected cost in efficacy, which prompted the addition of a third agent such as thiotepa, melphalan or clofarabine.
Total Body Irradiation Forever? Optimising Chemotherapeutic Options for Irradiation-Free Conditioning for Paediatric Acute Lymphoblastic Leukaemia. · Ben Hassine K, Powys M, Svec P, Pozdechova M, Versluys B, Ansari M, Shaw PJ · Frontiers in Pediatrics 2021What this evidence cannot show
A narrative review, cited as a history of how the regimen was assembled, not as evidence that any step in that history was correct. 'May not be as effective' is the reviewers' hedged summary of a comparison that was never made as a randomised trial in children; the substitution and the third-agent addition happened together in practice, so their separate effects cannot be recovered from the literature.
Evidence that argues against this (1)
- Whatever the assembly history, the finished regimen has now been compared head to head with irradiation and lost: in FORUM, 2-year overall survival was 0.91 with TBI plus etoposide versus 0.75 with fludarabine-based chemoconditioning, and relapse 0.12 versus 0.33. The toxicity-reduction rationale that produced fludarabine-containing conditioning was vindicated on non-relapse mortality and defeated on relapse. Peters C et al. 2021
the US treosulfan label mandates fludarabine 30 mg/m2/day on days -6 to -2 as its partner
BU/THIO/FLU, one of the two FORUM chemoconditioning backbones
fludarabine entered conditioning as a toxicity-reducing substitute for cyclophosphamide alongside busulfan, at a suspected cost in efficacy
lymphodepleting component of modern conditioning; absent from Interfant-06
third agent of BU/THIO/FLU and TREO/THIO/FLU, held constant in both FORUM chemoconditioning arms
the lymphodepleting agent of modern conditioning depletes the effector class any GVL effect must run through
two T-cell-suppressing agents in one regimen, one dosed per kilogram and one per square metre, neither monitored in Interfant-06
model-based dosing beat BSA dosing on graft rejection, 1.8% vs 8.5%
the three-species activation chain behind the single drug name