Vincristine - measured to be no different in KMT2A-rearranged disease, and measured to be UNDERDOSED in infants by the very rule meant to protect them
also called vincristine · VCR · vincristine sulfate · Oncovin · VCR sulphate · leurocristine · vinca alkaloid (class)
Vincristine stops cells dividing by wrecking the scaffolding that pulls chromosomes apart. Infant leukaemia cells are no more or less sensitive to it than any other child's - but infants are given a reduced dose on the assumption that they clear it slowly, and when someone finally measured, they did not.
The same thing, for a scientist
Vincristine is a vinca alkaloid that binds beta-tubulin at the vinca site and blocks microtubule polymerisation, arresting cells in metaphase. In Interfant-06 it is given **1.5 mg/m2/dose** IV push on induction days 8, 15, 22 and 29 and on OCTADA(D) part I days 1, 8, 15 and 22 [E1]. **It is the clearest 'in the protocol, and that is all' case in this beat, and the null is a measured one:** the infant in-vitro drug-resistance profile found no difference between infants and older children for vincristine, alongside no difference for anthracyclines, thiopurines or epipodophyllotoxins, in the same experiment that found >580-fold prednisolone resistance and 1.9-fold cytarabine sensitivity [E2]. The one KMT2A-relevant signal is lineage rather than fusion: 11q23-rearranged AML carries an 'AML profile' of relative resistance to glucocorticoids **and vincristine** [E3]. **The infant-specific finding is a dosing error, not a biology.** Vincristine clearance normalised to body surface area did not differ between neonates/infants and older children; the reduced doses customarily given to infants produced significantly lower AUC (P = 0.047), and doses below 0.05 mg/kg gave significantly lower exposure than doses at or above it (P <= 0.0001) [E4].
Where it sits, and the dose arithmetic that should worry a reader
Interfant-06 1: 1.5 mg/m2/dose IV push, four doses in induction (days 8, 15, 22, 29) and four in OCTADA(D) part I (days 1, 8, 15, 22).
Apply the protocol's own age rule - x 2/3 under 6 months - and an infant under 6 months receives an effective 1.0 mg/m2/dose. Barnett et al. characterised vincristine population PK across 57 childhood cancer patients including 26 neonates and infants, and found that BSA-normalised clearance did not differ between the infants and the older children, while the lower doses given to infants produced significantly lower AUC 4. Their recommendation is explicit: doses below 0.05 mg/kg should not be used in neonates and infants because of the risk of subtherapeutic exposure, and they propose a target window of 50-100 ug/L*h with therapeutic drug monitoring, which they showed to be feasible and well tolerated.
This is the most transferable finding in this beat. The 2/3 and 3/4 multipliers in Interfant-06 are stated on every chemotherapy page with no derivation and no reference 1. For the one drug where the assumption behind them has been tested directly in this age group, it was wrong, and wrong in the direction that undertreats a disease with 46.1% six-year EFS.
The KMT2A evidence: a well-controlled null, and one lineage signal
The null. Ramakers-van Woerden et al. ran 44 infants against 425 older children with common/preB ALL in one MTT experiment and reported: 'No differences were found for vincristine, anthracyclines, thiopurines, epipodophyllotoxines, or 4-hydroperoxy-ifosfamide' 2. This is a positive control-bearing null: the same assay, the same samples and the same comparison returned >580-fold for prednisolone and 12-fold for asparaginase, so the instrument demonstrably returns a difference when one exists. That is what distinguishes this from an unbounded 'nobody has looked'.
The one signal, and it is about lineage. Palle et al. describe an 'AML-profile' of relative resistance to glucocorticoids and vincristine shared by t(9;11) AML, other 11q23 AML and non-rearranged AML 3 - i.e. it travels with myeloid lineage, not with the rearrangement. In a disease that switches lineage under pressure, that is worth holding: vincristine is a drug the myeloid compartment is relatively resistant to, and the myeloid compartment is where KMT2A-rearranged infant ALL escapes to.
And the modest positive. Romidepsin synergised with vincristine in 4 of 6 infant ALL cell lines 5 - better than dexamethasone's 3/6, worse than daunorubicin's 6/6.
| target | beta-tubulin, vinca binding site - inhibits microtubule polymerisation and arrests cells in metaphase. No tubulin gene node exists in this graph as of cycle 28. |
|---|---|
| modality | small molecule, vinca alkaloid, microtubule-destabilising agent |
| interfant06 position | remission induction days 8, 15, 22, 29; OCTADA(D) part I days 1, 8, 15, 22 |
| interfant06 dose | 1.5 mg/m2/dose IV push |
| infant dose adjustment | x 2/3 under 6 months (effective 1.0 mg/m2/dose), x 3/4 from 6 through 12 months |
| kmt2a specific evidence | A MEASURED NULL. No difference between infants and older children in the MTT assay, in the same experiment that returned >580-fold for prednisolone - so this is a controlled negative, not an untested assumption. The only signal is lineage-level relative resistance in 11q23 AML. |
| infant specific finding | BSA-normalised clearance is NOT different in neonates/infants; the customary dose reduction produces significantly lower AUC; doses <0.05 mg/kg should not be used; proposed therapeutic window 50-100 ug/L*h with feasible TDM |
| protocol toxicity rules | reduce to 2/3 for severe paresis or constipation and re-escalate if symptoms resolve; withhold for foot drop or ileus until fully resolved then restart at 2/3; do NOT modify the dose for jaw pain, use analgesics |
key ae
[ "peripheral sensorimotor neuropathy", "autonomic neuropathy with constipation and ileus", "jaw pain", "foot drop", "SIADH", "severe local tissue necrosis on extravasation", "fatal if given intrathecally" ]
resistance
[ "no infant- or MLL-specific mechanism identified in this beat", "relative resistance as part of a myeloid-lineage profile in 11q23-rearranged AML - relevant because this disease switches lineage", "classical P-glycoprotein efflux, which was NOT elevated in infants relative to older children in the cited profile study" ]
trials
[ "NCT00550992" ]
- **The dose rule intended to protect infants has been shown, for this drug, to underdose them.** Correcting it would raise exposure to a neurotoxin in patients who cannot report neuropathy, in a protocol where vincristine is given eight times. The safe-looking move and the effective-looking move point in opposite directions and neither has outcome data.
- Vincristine is fatal if given by the intrathecal route, and this protocol gives intrathecal therapy at eleven or more timepoints in the same patients. That is a systems risk, not a pharmacological one, and it is the reason vincristine is never dispensed in a syringe in units that also give intrathecal drugs.
- The in-vitro null means there is no biological argument for keeping vincristine in a future KMT2A-directed regimen - and equally none for removing it, since the null is against older children rather than against placebo. A drug with a measured null is the easiest thing to drop and the hardest to justify dropping.
- If lineage switch confers the myeloid vincristine-resistance profile, then vincristine is contributing least exactly at the point of relapse it would most need to cover, and a regimen that leans on it during reinduction may be leaning on the wrong agent for the cells that survived.
- Vincristine is a CYP3A4 substrate and infants in induction receive azoles for fungal prophylaxis, which inhibit CYP3A4. The protocol's fixed dose reduction interacts with an unmeasured and highly variable drug-drug interaction in the opposite direction.
- Is the 2/3 dose multiplier under 6 months costing leukaemia control? For vincristine the clearance assumption behind it has been tested and failed, and the exposure shortfall is measured. Nobody has linked infant vincristine AUC to event-free survival.
- What is vincristine's dose-limiting toxicity in a 4-month-old, and how would anyone know? Peripheral neuropathy is assessed by symptoms and by examination findings a preverbal infant cannot supply, so the toxicity that sets the dose in older children is effectively unmeasurable here.
- Does a KMT2A-rearranged blast that switches to a myeloid phenotype acquire the vincristine resistance that characterises myeloid lineage? The lineage-level difference is measured; the within-patient version never has been.
- Would therapeutic drug monitoring change anything in infant ALL? It was shown feasible and well tolerated in neonates and infants with mixed cancers, and has never been tried in an infant leukaemia protocol.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Interfant-06 gives vincristine 1.5 mg/m2/dose by IV push on induction days 8, 15, 22 and 29 and on OCTADA(D) part I days 1, 8, 15 and 22, with the body-surface-area dose multiplied by 2/3 under 6 months and 3/4 from 6 through 12 months; the protocol directs reduction to 2/3 of the recommended dose for severe paresis or constipation, withholding for foot drop or ileus with re-introduction at 2/3, and explicitly no dose modification for jaw pain.
Interfant-06 treatment protocol for infants under 1 year of age with acute lymphoblastic leukaemia - International collaborative treatment protocol, SKION version 01 (April 2006): chapter 5 chemotherapy schedule and chapter 7 modification for toxicity · Interfant-06 study group / SKION (Stichting Kinderoncologie Nederland, Dutch Childhood Oncology Group) · Interfant-06 study protocol document (90 pp.) 2006What this evidence cannot show
A trial protocol document, not a peer-reviewed publication, and specifically the SKION (Dutch) national version of April 2006; the trial was amended during its course (asparaginase provisions in MARMA and OCTADA(D) among others), so later versions may differ. Doses given are PLANNED doses; delivered cumulative dose per patient is lower wherever a course was omitted, delayed or curtailed by a stopping rule, and the trial report does not publish delivered dose per patient. Retrieved and text-extracted by this beat on 2026-09-20 (90 pages, 194,979 characters); page numbers are the PDF's own. The toxicity-modification rules are from chapter 7 and are recommendations to the treating centre, not measured thresholds; the protocol gives no evidence base for the 2/3 toxicity reduction either.
Evidence that argues against this (1)
- The protocol contains no maximum absolute dose cap for vincristine, which is the conventional 2 mg ceiling in older children. Its absence is unremarkable at infant body surface areas, but it means the protocol's vincristine provisions cannot be read across to the older children in whom most vincristine neurotoxicity data were generated. Barnett S et al. 2022
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In an MTT assay comparing 44 infants with ALL against 425 children over 1 year with common/preB ALL, no difference in cellular drug resistance was found for vincristine, anthracyclines, thiopurines, epipodophyllotoxins or 4-hydroperoxy-ifosfamide, in the same experiment that found infants more than 580-fold resistant to prednisolone, 12-fold resistant to L-asparaginase and 1.9-fold more sensitive to cytarabine.
In vitro drug-resistance profile in infant acute lymphoblastic leukemia in relation to age, MLL rearrangements and immunophenotype. · Ramakers-van Woerden NL, Beverloo HB, Veerman AJ, Camitta BM, Loonen AH, van Wering ER, et al. · Leukemia 2004What this evidence cannot show
A null from an assay that returned large differences for three other drugs in the same samples, which is the strongest form this kind of negative can take. It remains an in-vitro LC50 on bulk primary blasts after short-term culture: it cannot detect a subclonal difference, a stromal-dependent difference, or a difference in a pharmacokinetic rather than cellular variable. No confidence interval is given for the vincristine comparison in the abstract, so the bound on the null effect size is unknown.
Evidence that argues against this (1)
- The null is about infants versus older children with a different immunophenotype, not about KMT2A rearrangement specifically. An independent study using a different assay found relative vincristine resistance in 11q23-rearranged AML as part of a myeloid-lineage profile, so a lineage-level vincristine difference exists in the wider KMT2A-rearranged population even though the infant-ALL comparison is null. Palle J et al. 2005
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In a fluorometric microculture cytotoxicity assay across 132 children with AML and 178 with ALL, t(9;11) AML, other 11q23-rearranged AML and non-rearranged AML shared an 'AML profile' of relative resistance to glucocorticoids and vincristine.
Cellular drug sensitivity in MLL-rearranged childhood acute leukaemia is correlated to partner genes and cell lineage. · Palle J, Frost BM, Forestier E, Gustafsson G, Nygren P, Hellebostad M, et al. · Br J Haematol 2005What this evidence cannot show
Subgroups are small (n = 10 for t(9;11), n = 14 for other 11q23) and the profile is described qualitatively as 'common to all three groups' rather than with a fold-difference or a P value for vincristine in the abstract. Because the profile is shared with non-rearranged AML, it is a lineage statement and carries no information about the rearrangement.
Evidence that argues against this (1)
- A lineage-associated resistance measured in AML samples does not establish that a lymphoid KMT2A-rearranged blast becomes vincristine-resistant when it switches lineage, which is the clinically interesting version of the claim. Nobody has measured drug sensitivity across a lineage switch in the same patient. Ramakers-van Woerden NL et al. 2004
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In a population pharmacokinetic analysis of 57 childhood cancer patients including 26 neonates and infants, vincristine clearance normalised for body surface area did NOT differ significantly between neonates/infants and older children; the lower doses given to neonates and infants (0.02-0.05 mg/kg or 0.75-1.5 mg/m2) produced significantly lower AUC than in older children receiving 1.5 mg/m2 (P = 0.047), doses below 0.05 mg/kg gave significantly lower AUC than doses at or above 0.05 mg/kg (P <= 0.0001), and the authors conclude that doses under 0.05 mg/kg should not be used in this population, proposing a therapeutic window of 50-100 ug/L*h with therapeutic drug monitoring shown to be feasible and well tolerated.
Vincristine dosing, drug exposure and therapeutic drug monitoring in neonate and infant cancer patients. · Barnett S, Hellmann F, Parke E, Makin G, Tweddle DA, Osborne C, et al. · Eur J Cancer 2022What this evidence cannot show
57 patients across mixed childhood cancers, not an infant ALL cohort, and the exposure-response relationship for antileukaemic effect is not established - lower AUC is shown, inferior outcome is not. The proposed 50-100 ug/L*h window is derived, not validated against efficacy. Vincristine is also a CYP3A4 substrate whose ontogeny in the first year is itself a moving target, so a single clearance comparison does not settle developmental pharmacology.
Evidence that argues against this (1)
- Infant dose reduction was introduced to reduce toxicity, not to match clearance, and vincristine's dose-limiting toxicity is peripheral and autonomic neuropathy that an infant cannot report. Restoring full BSA dosing on the strength of an exposure argument would raise a toxicity that this cohort is structurally unable to detect early, and the Barnett data do not measure neurotoxicity as a function of exposure. Interfant-06 study group / SKION (Stichting Kinderoncologie Nederland et al. 2006
- E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Romidepsin showed synergy with vincristine in four of six infant ALL cell lines, against six of six for daunorubicin and 4-hydroperoxycyclophosphamide, three of six for dexamethasone, and antagonism in four of six for methotrexate and L-asparaginase.
Romidepsin enhances the efficacy of cytarabine in vivo, revealing histone deacetylase inhibition as a promising therapeutic strategy for KMT2A-rearranged infant acute lymphoblastic leukemia. · Cheung LC, Cruickshank MN, Hughes AM, Singh S, Chua GA, Ford J, et al. · Haematologica 2019What this evidence cannot show
Read from the PMC HTML full text on 2026-09-20 (the efetch and Europe PMC abstract fields are empty because this is a Haematologica letter). Six cell lines, Bliss-independence scoring, no primary samples, no in-vivo arm for this pair.
Evidence that argues against this (1)
- The in-vivo arm of the same paper tested only cytarabine, so the vincristine synergy score has no animal confirmation. Cell-line synergy scores predict clinical combination benefit poorly, and a 4/6 result in a panel whose best agent scores 6/6 is a middling signal. Cheung LC et al. 2019
- part of →Interfant-06 — 651 infants, and the definitive answer that myeloid-style consolidation does not help
1.5 mg/m2 IV push, induction days 8/15/22/29 and OCTADA(D) days 1/8/15/22
induction and reinduction
in the protocol; in-vitro sensitivity indistinguishable from older children's
the measured null and the measured extreme from the same MTT experiment on the same samples
synergy in 4 of 6 infant ALL cell lines
both measured null against older children in the same MTT experiment