1. The Claim, Restated Honestly

In September 2022 the United States Food and Drug Administration approved daxibotulinumtoxinA-lanm (Daxxify) for the temporary improvement of moderate to severe glabellar lines. The launch was accompanied by a duration claim that has since defined the product's commercial identity: a median effect of approximately 24 weeks — roughly twice the 12–16 weeks conventionally quoted for onabotulinumtoxinA (Botox), abobotulinumtoxinA (Dysport) and their peers. Marketing copy has, at times, gone further, implying a categorical pharmacological superiority as though a fundamentally different kind of toxin had entered the clinic.

The claim deserves respect but not deference. Daxxify is a competently developed product from a serious company, and the SAKURA phase III programme is a well-conducted set of trials. But the mechanistic implication — that a toxin can be "more irreversible" than another, or can somehow escape the biological ceiling imposed by nerve regeneration — is not supported by the underlying neuromuscular pharmacology. This editorial is not an attack on Daxxify. It is a request that we describe it accurately.

2. First Principles: Botulinum Toxin Is Already Effectively Irreversible

The pharmacology of botulinum neurotoxin type A is one of the best-characterised in modern medicine (Dressler 2012; Pirazzini et al. 2017). The 150 kDa dichain protein binds via its heavy chain to synaptic vesicle protein SV2 and the polysialoganglioside GT1b on the presynaptic membrane of the cholinergic motor nerve terminal. Following receptor-mediated endocytosis and pH-dependent translocation of the light chain into the cytosol, the zinc-dependent metalloprotease activity of the light chain cleaves the 25 kDa synaptosomal-associated protein SNAP-25 between residues 197 and 198. The cleaved SNAP-25 is unable to participate in SNARE-complex assembly, and the vesicle fusion machinery of that terminal is disabled. The nerve terminal is, in functional terms, silent.

Nothing about that silence is "reversible" in the conventional pharmacological sense. The toxin does not dissociate from a receptor to allow the muscle to move again. The cleaved SNAP-25 is not repaired. The terminal itself does not spontaneously regenerate its lost SNARE machinery. Clinical recovery, when it comes, is a completely separate biological process: the neuron responds to functional denervation by producing collateral axonal sprouts (de Paiva et al. 1999), the sprouts form new, immature neuromuscular junctions on the muscle fibre, and — over weeks to months — the original terminal is remodelled and the sprouts are pruned back. The end-organ is re-innervated. The clinical effect wears off.

This is the crucial mechanistic point. The duration of effect of any botulinum toxin A product is not determined by how long the toxin "lasts" at the receptor. It is determined by how long the nerve takes to grow around the blockade. Any claim that a new toxin is longer-lasting because it is more stubbornly bound must, in principle, be reconciled with the fact that the recovery clock is a neuronal clock, not a pharmacological one.

3. What Daxxify Actually Is

DaxibotulinumtoxinA-lanm is a purified 150 kDa botulinum neurotoxin type A — the same molecular species as onabotulinumtoxinA and incobotulinumtoxinA (Carruthers et al. 2020). It differs from Botox in two formulation-level respects. First, it contains no human serum albumin as an excipient. Second, it is stabilised by a proprietary positively charged synthetic peptide, RTP004, described by Revance as a "peptide excipient" that binds electrostatically to the toxin.

This is not a new toxin serotype. It is not a modified neurotoxin. The light chain cleaves SNAP-25 at the same 197–198 bond. The heavy chain binds the same SV2/GT1b complex on the same cholinergic terminal. If the peptide is doing something clinically important, it is doing it before or during nerve entry — not after.

4. RTP004 and the Delivery Hypothesis

Revance's original scientific communications described RTP004 as a cell-penetrating peptide-like excipient that increases the surface association of the toxin with target cell membranes and its subsequent uptake (Waugh et al. 2020). Later peer-reviewed work has broadly supported the idea that RTP004 improves neuronal binding and intracellular delivery of the toxin in experimental preparations. Reviews of the SAKURA programme have consistently framed the peptide as a delivery-enhancement excipient rather than as an agent that alters the intracellular fate of the toxin (Bertucci & Solish 2023).

This wording matters. What the pre-clinical literature supports is that RTP004 helps more toxin molecules reach the inside of more nerve terminals. What it does not support is any claim that RTP004 makes the toxin, once inside the terminal, longer-lived. There is no published evidence that RTP004 stabilises the light chain against intracellular turnover, that it inhibits axonal sprouting, that it interferes with junctional remodelling, or that it changes the biology of nerve recovery in any way. If it did, we would expect analogous findings in the cervical dystonia programme, where recovery kinetics would be an obvious secondary endpoint. We do not have them.

5. The SAKURA Trials: Reading the Numbers Carefully

SAKURA-1 and SAKURA-2 were the two pivotal randomised, double-blind, placebo-controlled phase III trials of daxibotulinumtoxinA for glabellar lines. SAKURA-3 was the open-label long-term safety extension. Across the programme, a single 40 U treatment produced a median duration of effect on the Investigator Global Assessment–Frown Wrinkle Severity scale of approximately 24 weeks, with a subset of subjects maintaining a response beyond 27 weeks and some out to 36 weeks (Carruthers et al. 2020; Bass et al. 2021).

Two features of the design deserve emphasis. First, the dose. Forty units of daxibotulinumtoxinA is not directly interconvertible with 20 units of onabotulinumtoxinA. Botulinum toxin unit potency is assay-specific; each manufacturer's LD50 mouse bioassay is calibrated separately, and no international unit exists. The comparison "40 U Dax versus 20 U Ona" is a comparison of two different measuring sticks. The relevant clinical exposure question is not units but biological effect, and the biological effect at the injected doses used in SAKURA is substantial — a very deep chemodenervation of the glabellar complex.

Second, the endpoint. SAKURA measured time to loss of response, where response was defined as at least a two-point improvement from baseline. A very deep chemodenervation takes longer to "drop below the two-point line" simply because it has further to fall. This is not a trivial artefact — it is arguably the whole story. If you produce a deeper initial denervation, the recovery curve — which is set by nerve sprouting — begins from a lower floor and takes longer to cross any predefined severity threshold.

The comparator arms are also relevant. Head-to-head data against Botox are not part of the pivotal glabellar programme. The interpretive burden — Daxxify "lasts twice as long as Botox" — is being carried by cross-trial comparison against historical duration figures for onabotulinumtoxinA. Cross-trial comparison in toxin pharmacology has a long history of producing figures that do not survive contact with prospective randomised head-to-head data (Frevert 2015; Kane et al. 2021).

6. The Cervical Dystonia Signal

The ASPEN-1 and ASPEN-OLS trials evaluated daxibotulinumtoxinA in cervical dystonia, a therapeutic indication in which duration of effect is measured against a very different clinical yardstick — the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS). In this population, the median duration was approximately 24 weeks at the higher dose tested (Jankovic et al. 2023). This is a genuinely useful signal for therapeutic patients, whose retreatment intervals with existing toxins are frequently limited by symptom breakthrough at 10–12 weeks.

But even here, the extended duration is compatible with the deep-denervation hypothesis. The doses used are large. The muscles are large. The therapeutic ceiling being pressed against is the same neuronal recovery ceiling. Nothing in the ASPEN data set requires a novel mechanism. It requires more toxin, more efficiently delivered, into more nerve terminals.

7. The Dose Equivalence Argument

If duration scales primarily with the depth of the initial chemodenervation, then a straightforward corollary follows: an experienced injector using an existing product can, within safety limits, produce a longer duration of effect by placing a modestly larger dose more precisely into the motor endplate zone. This is not a novel observation. It has been the tacit practice of expert injectors for years. Comparative work with onabotulinumtoxinA has repeatedly shown a dose-duration relationship in the glabella: 30–40 U produces a longer effect than 20 U, at the cost of a heavier brow-drop risk profile and — critically — a heavier appearance that is not always what the patient asked for (Carruthers et al. 2005; Rzany et al. 2013).

The Daxxify duration figure is not a magical property of the molecule. It is, in large part, a property of the product being licensed and marketed at a dose that produces a very deep glabellar denervation. Any experienced practitioner willing to press their existing molecule to a comparable depth of effect — and willing to accept the same trade-offs in expressive range and brow position — can shorten the gap substantially. The trade-offs, and the aesthetic judgement about how much movement to leave in the upper face, are the actual clinical craft. The molecule is downstream of the craft.

8. The Learning Curve Nobody Costs In

There is a second, quieter argument against uncritical adoption, and it is the one that matters most for readers of this journal. Every botulinum toxin brand has its own diffusion behaviour, its own reconstitution habits, its own onset kinetics, its own perceived spread, its own idiosyncratic behaviour in the frontalis, the crow's feet, the depressor anguli oris and the masseter. A practitioner who has spent a decade with onabotulinumtoxinA has, whether they articulate it or not, an internal predictive model of that molecule. They know how a fractional-unit adjustment in the corrugator will trend the medial brow. They know how the lateral orbicularis oculi behaves at week two versus week six in a 47-year-old with a thin, mobile brow. They know when to underdose because they know what over-treatment looks like in this specific molecule on this specific face.

That internal model is not transferable. It is a molecule-specific learning curve, accumulated across thousands of cases and — critically — across the small number of unsatisfying outcomes that recalibrated their internal model. The Dunning–Kruger commentary published elsewhere in this volume argues that structured supervised exposure is the mechanism by which injectors move from the early confidence peak into calibrated competence. Switching molecules resets a portion of that calibration. It does not reset it to zero — the underlying anatomy is unchanged — but it re-opens a window of miscalibration that the practitioner had, at some cost, closed.

The commercial framing of "just switch" hides this cost. It is presented as a menu change. It is, in reality, a partial re-entry into the early learning curve of a new molecule at the exact moment when the practitioner is being told to charge the patient a premium price for the switch. The economics reward the manufacturer; the risk is carried by the patient and, eventually, by the practitioner.

9. What a Fair Comparison Would Look Like

The intellectually honest comparison that the field needs, and which has not yet been done at scale, is a prospective, randomised, split-face or split-cohort head-to-head trial of daxibotulinumtoxinA against a dose-optimised onabotulinumtoxinA arm, with duration measured against a matched threshold and with practitioner-blinded injection. Absent that, the field is being asked to draw conclusions from cross-trial comparisons in which the daxibotulinumtoxinA arm is dosed to the ceiling of its therapeutic window and the historical Botox comparator is dosed to a conservative label-recommended value that most expert injectors have not used in years.

Head-to-head data of that quality would settle the mechanistic question quickly. If daxibotulinumtoxinA still outperformed a dose-matched Botox arm by six to eight weeks, the RTP004 delivery hypothesis would gain real strength. If the gap collapsed to two to three weeks — as the deep-denervation hypothesis predicts — the field would have a much more sober conversation about whether a molecule-switch, a retraining cost and a price premium are justified for a benefit that a competent injector can largely reproduce with the molecule already on their tray.

10. The Editorial Position

Aesthetic Intelligence takes the following editorial position on the Daxxify duration claim.

First, we accept that daxibotulinumtoxinA, at the licensed glabellar dose, produces a clinically longer duration of effect than a conservatively dosed onabotulinumtoxinA arm in cross-trial comparison. That is what the SAKURA data show.

Second, we reject the mechanistic framing that this is because Daxxify is "more irreversible" or has escaped the recovery ceiling imposed by neuronal sprouting. Both molecules cleave SNAP-25 in the same essentially irreversible manner, and recovery in both is governed by the same neurobiology. The most parsimonious explanation for the duration gap is a deeper initial chemodenervation, driven by more efficient membrane binding and intracellular delivery of the toxin via the RTP004 excipient. This is a delivery advantage, not a persistence advantage.

Third, we note that a substantial fraction of the duration gap is reproducible by an experienced injector willing to place a modestly larger dose of an existing toxin more precisely into the motor endplate zone. The trade-offs — brow position, expressive range, patient-reported "heaviness" — are the same trade-offs that experienced injectors already navigate every day. This is craft, not chemistry.

Fourth, we caution that switching molecules is not a free operation. It resets part of a molecule-specific learning curve that took years to build, at the point in a practitioner's career when their internal predictive model of the existing molecule is at its most valuable. That cost is systematically excluded from the marketing case, and it is the cost most likely to be paid, silently, by the patient.

Fifth, and finally, we welcome Daxxify as a legitimate addition to the botulinum toxin armamentarium — particularly for therapeutic indications such as cervical dystonia where a longer inter-dose interval carries real quality-of-life value — while insisting that its aesthetic marketing be brought into line with its actual pharmacology. It is a well-formulated toxin. It is not a different biology.

11. Conclusion

The provocative claim that a botulinum toxin can be "more irreversible" than another collapses on inspection. Both Daxxify and Botox produce essentially permanent cleavage of SNAP-25 in the affected terminal. Both rely on axonal sprouting for clinical recovery. The duration difference visible in cross-trial comparison is best explained by a deeper initial denervation driven by improved delivery, not by a novel resistance to neuronal recovery. That distinction matters, because it changes the clinical question from "Do I need to switch to keep up?" to "Is the retraining cost of switching worth a benefit I can largely reproduce with a molecule I have spent a decade learning to place?"

For most experienced injectors, on most patients, the answer is no. For a specific subset of patients — those who genuinely value a longer inter-treatment interval and who accept the aesthetic and financial trade-offs that come with a deeper denervation — Daxxify is a rational choice. The rest is marketing.