Insight
Erythroid data exist for two transferrin receptor programmes and no others
The transferrin receptor is the iron-uptake receptor on erythroid precursors, so a shuttle that binds it is binding something with a day job. Two clinical programmes have disclosed figures on that. Every other programme in the class has disclosed nothing, and the gap is more informative than either figure.
What has been disclosed
| Programme | Finding as disclosed | n / N |
|---|---|---|
| tividenofusp alfa, Denali | Anaemia, with its own Warnings and Precautions section | 24 / 47 |
| trontinemab, Roche, extension, previously placebo | Anaemia, CTCAE v5.0, clinically significant per investigator | 1 / 22 |
| trontinemab, Roche, extension, Q4W induction | As above | 1 / 23 |
| trontinemab, Roche, extension, Q12W maintenance | As above | 3 / 81 |
| pabinafusp alfa and the other JCR assets | No anaemia or haemoglobin figure located | not disclosed |
| ABL301, ABL Bio | No anaemia or haemoglobin figure located | not disclosed |
| ALIA-1758, AbbVie | No Phase 1 safety data disclosed at all | not disclosed |
| verenafusp alfa, AO GENERIUM | No safety data located | not disclosed |
AVLAYAH US Prescribing Information, revised 3/2026, Reference ID 5769209, section 5.3; Kulic et al., AAIC 2026, slide 18, open-label extension Weeks 1 to 74, snapshot 5 September 2025.
The regulator took it seriously
Anaemia in the tividenofusp alfa label is not a line in an adverse event table. It carries a dedicated Warnings and Precautions section, haemoglobin monitoring at baseline and at three months, and a management note stating that treatment may include supplementation with iron.
AVLAYAH US Prescribing Information, revised 3/2026, Reference ID 5769209, section 5.3.
That is a regulator-level document reporting the finding in one transferrin receptor asset, and a conference presentation reporting the same category of finding in a second, structurally different one. Two independent observations in the same receptor class is worth noticing.
The two figures are not comparable
Setting 51% against 4.5% would be the obvious move and it would be wrong. The denominators describe different things.
The tividenofusp alfa denominator of 47 is a paediatric MPS II cohort in a Phase 1/2 study, on chronic enzyme replacement, in a population with its own baseline haematology.
ClinicalTrials.gov NCT04251026.
The trontinemab figures cover the open-label extension only, Weeks 1 to 74, in adults with Alzheimer's disease. They are not the incidence across the double-blind parts of the study, which has not been reported in any document retrieved here.
Kulic et al., AAIC 2026, slides 17 and 18.
Different populations, different exposures, different durations, different assays for what counts as an event. No rate in the table above should be set against another one, and the table is grouped by receptor rather than ranked for that reason.
A figure that was struck
A secondary review reported that up to 18% of trontinemab participants developed transient anaemia. The AAIC 2026 presentation reports 4.5%, 4.3% and 3.7% across its three extension groups. No document retrieved supports 18%.
The presentation covers the extension only, so it does not formally refute a figure drawn from elsewhere in the study. It also does not support one. The 18% was struck and the strike recorded, which is a different thing from deleting it.
Nature Reviews Drug Discovery news feature, 16 July 2026, for the 18% figure. Kulic et al., AAIC 2026, slide 18, for the extension rates.
Preclinical platforms are designing around it
Several preclinical shuttle platforms describe engineering intended to spare erythroid cells. One carrier is described as designed to spare reticulocytes. Another platform describes Fc engineering to reduce reticulocyte loss, on a shuttle said to bind the same epitope as trontinemab.
Both claims rest on a single secondary review with no company primary document behind either, so both sit at LOW confidence and are struck from every count in the underlying market map. They are mentioned here only because a design choice is itself evidence of what its designers expect.
Nature Reviews Drug Discovery news feature, 16 July 2026. No company primary document retrieved for either platform.
What this table is not
It is not a class safety claim. Two programmes disclosing a finding is not the same as a class effect, and six programmes disclosing nothing is not evidence of absence.
There is also no comparator here. Nothing in the underlying file establishes what the anaemia rate looks like in a non-transferrin-receptor delivery class, because no such comparison is supported by the sources. Whether the mechanism explains the finding is an inference from pharmacology, and it is carried in the file as an inference rather than as a mechanism claim.
The finding that stands without qualification is the disclosure gap. Nine programmes are covered by that table and seven of them have published no erythroid data at all, including two that have completed Phase 1.