Nobody needs to enrol anyone: the PTH question is a regression on data that already exists
also called PTH covariate reanalysis revumenib · paediatric subgroup endocrine labs menin
The key question - is the parathyroid hormone rise caused by the drug or by the kidneys - can be answered with the blood results already sitting in the sponsor's database, using statistics rather than new patients.
The same thing, for a scientist
An analysis, not a study. The revumenib registration dataset already contains, for 241 patients including 34 children, serial post-baseline intact PTH, phosphate, corrected calcium, creatinine and alkaline phosphatase, plus individual drug exposure from the PK programme [E1]. Three questions follow directly and need no new subject: (1) does intact PTH rise with drug exposure after adjustment for creatinine and phosphate; (2) what are the endocrine laboratory rates in the 34 paediatric patients separately from the 207 adults; (3) does PTH fall in patients who discontinue and have a post-treatment value. All three are currently unanswerable from public sources only because the label reports pooled marginal rates.
The three models
M1 - exposure dependence. Mixed-effects model of log(intact PTH) over time with individual exposure (AUC or trough) as the predictor and creatinine, phosphate, albumin-corrected calcium and, if available, 25-OH vitamin D as covariates. An exposure coefficient surviving adjustment is the on-target signature; a coefficient that vanishes when creatinine enters is the renal null.
M2 - paediatric stratification. The same laboratory rates as the label's Table 8, computed separately for the 25 infants under 2, the 78 children aged 2 to under 12, the 29 adolescents and the adults. A higher paediatric rate would be the first human evidence that a developing endocrine system is more sensitive.
M3 - reversibility. Among patients with a post-discontinuation value, the trajectory of PTH, phosphate and ALP off drug. This is the single most informative number in the dataset and does not appear anywhere in public.
Status and how to run it
Status: planned, and not runnable by this project alone. The individual-level data are not public. Routes: (a) the FDA multi-discipline review for NDA 218944, which is public and may contain exposure-response analyses of laboratory parameters that never reached the label - this has not been read and is the cheapest next step; (b) a data-sharing request to the sponsor; (c) inclusion of these three models in the analysis plan of PMR 4730-1, which FDA has already required for exactly this question.
What this project can do now is state the models precisely enough that whoever holds the data can run them, and record that the questions are answerable without a new trial.
| script | analyses/menin-pth-exposure-response/run.py |
|---|---|
| status | planned |
| result summary | Three prespecified models: exposure-response for intact PTH adjusted for creatinine and phosphate; paediatric stratification of endocrine laboratory rates; post-discontinuation trajectory. None requires a new subject. |
inputs
[ "revumenib registration laboratory dataset (not public)", "revumenib PK exposure dataset (not public)", "FDA multi-discipline review NDA 218944 (public, unread)" ]
- If the exposure coefficient vanishes on adjustment, a substantial part of this beat's parathyroid argument is withdrawn, and the analysis should be run by people prepared for that outcome.
- The paediatric stratification could show the opposite of what this beat expects - lower rates in children - and that would be equally publishable and equally important.
- Relying on an analysis nobody outside the sponsor can run means the graph's central parathyroid claim is, for now, unfalsifiable from public data.
- Does the FDA multi-discipline review for NDA 218944 contain an exposure-response analysis for intact PTH? Reading it is free and has not been done.
- How many patients in the registration dataset have any post-discontinuation endocrine laboratory value at all?
- Would the sponsor accept an academic data-sharing request framed around a paediatric safety question?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
The revumenib label reports pooled intact PTH (34%), phosphate (51% up, 25% down), creatinine (38%), ALP (33%) and corrected calcium (15%) for 241 patients including 34 children, with no exposure-response, no paediatric stratification and no post-treatment values.
REVUFORJ (revumenib) tablets, for oral use - US Prescribing Information (SPL version 5, published 20 Feb 2026) · Syndax Pharmaceuticals, Inc. · DailyMed / FDA Structured Product Label, setid 6eb3cdbc-0e74-477d-82d6-3bb172d3f63f 2026What this evidence cannot show
The label is a summary. The absence of these analyses from the label does not mean they were not performed; the FDA review may contain them and has not been read for this entry.
Evidence that argues against this (1)
- FAERS contains zero endocrine or bone-mineral reports for revumenib against 9.68 expected, which shows spontaneous reporting cannot substitute for this analysis - and is why FDA required a trial instead. See ana-faers-menin-inhibitor-endocrine-disproportionality. Syndax Pharmaceuticals et al. 2026
- E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Menin is required for TGF-beta restraint of PTH secretion, giving the on-target hypothesis a mechanism that predicts exposure-dependence rather than creatinine-dependence.
Menin inactivation leads to loss of transforming growth factor beta inhibition of parathyroid cell proliferation and parathyroid hormone secretion. · Sowa H, Kaji H, Kitazawa R, Kitazawa S, Tsukamoto T, Yano S, et al. · Cancer Research 2004What this evidence cannot show
Ex vivo human parathyroid cells; antisense knockdown, not drug.
Evidence that argues against this (1)
- No parathyroid signal appears on the ziftomenib label, so a class-wide exposure-dependent parathyroid effect is not corroborated. Kura Oncology et al. 2025
- ← related toRUNX1::RUNX1T1 (AML1-ETO), t(8;21)(q22;q22)
Subgroup replication failure (Qin 2014, PMID 25449688): in 351 CBF-AML patients c-KIT mutation was adverse only in ADULT t(8;21) and had no significant effect in paediatric t(8;21) (p>0.05), undercutting the 129-patient c-KIT/transplant subgroup.
- ← related toNUP98::NSD1, t(5;11)(q35;p15.5)
Covariate correction (Ostronoff 2014, PMID 25145343): in 1421 patients NUP98/NSD1 was 7% of CN-AML, not 16.1%, and NUP98/NSD1 WITHOUT FLT3-ITD lost its prognostic significance in multivariable analysis.
P2: does the PTH rise track drug exposure after adjustment for creatinine and phosphate? Answerable today on existing data
Three models on data that already exists: exposure-response adjusted for creatinine and phosphate, paediatric stratification, and post-discontinuation trajectory
These three models belong in the analysis plan of PMR 4730-1, which FDA required for exactly this question
- tests →What the labels measure in a treated child: bone growth, in one sentence, and nothing else endocrine
Model M2 would produce the paediatric stratification whose absence that entry documents