A one-time cure has a commercial shape no chronic drug shares. It sells to the patients who already carry the diagnosis, and once that backlog is treated, it sells to almost no one. Novartis chief executive Vas Narasimhan described the mechanism plainly on an earnings call, explaining why sales of the spinal muscular atrophy gene therapy Zolgensma had stalled: as new markets open, the company penetrates a bolus of prevalent patients and then moves back to the incident population. For a forecaster raised on recurring-revenue models, that shape is the whole problem.
A one-time cure sells to a backlog, then a trickle
Zolgensma is the clearest case because it is the flagship. Net sales ran to roughly $920 million in 2020 and $1.35 billion in 2021 as the prevalent pool of diagnosed children was treated, then inverted to $311 million in 2023 and $295 million in 2024 on Novartis' filings. Nothing about the therapy changed. The backlog simply ran out, leaving the thin flow of newly diagnosed infants that newborn screening feeds each year. A prevalent backlog is a stock you deplete once; the incident population is a flow you cannot accelerate.
Approval was the easy part
The pattern is not confined to one asset. bluebird bio withdrew Zynteglo and Skysona from Europe in 2021 after payers declined to cover a roughly $1.8 million one-time price. BioMarin's hemophilia A therapy Roctavian, listed near $2.9 million, treated five patients and booked $7 million in early 2024 before the company narrowed its markets and ultimately withdrew it. Pfizer discontinued its approved hemophilia B gene therapy Beqvez in early 2025, less than a year after launch, with essentially no commercial patients. In each case the therapy worked and the market did not, because a durable clinical result met a payer system built to pay for chronic treatment over years, not a single high-cost event.
Capacity, not demand, caps uptake
Even where demand is not the issue, uptake is capped upstream of it. More than two years after approval, Vertex's sickle cell therapy Casgevy had reached only about 60 patients across the United States, Europe and the Middle East, with clinicians naming apheresis, the ability to collect enough of a patient's stem cells, as the binding bottleneck rather than any shortage of eligible patients. A chronic-drug forecast assumes eligible patients convert at a rate set by access and adoption; for a one-time therapy, the rate is set by how many procedures a finite network of centres can physically run in a year.
You model a depleting bolus, not an annuity
A defensible forecast for a one-time therapy has two moving parts, not one prevalence figure: the size of the prevalent backlog, and the cadence at which a new identified pool refills it. That second variable is the one the chronic model has no place for, and it is decisive. Zolgensma is now expected to grow again on a new formulation reaching older patients and continued ex-US newborn-screening rollout, each a fresh bolus drawn from a newly identified pool. Gene-therapy revenue is therefore not a single depleting curve but a series of them, and what determines the next one is not efficacy or price, fixed at approval, but the rate at which the system identifies the next cohort.
The forecast driver is identification, not demand
Read as an annuity, a gene therapy looks like a failing brand the moment its backlog clears. Read as a depleting bolus refilled by identification, it looks like what it is. This is one of the shifts behind the rare-disease sizing discipline: for a one-time therapy you forecast the backlog, the depletion, and the cadence that refills the pool, not a straight-line penetration of prevalence.
AXLRx models one-time therapies as a depleting backlog plus an identification-driven refill, with treatment-centre capacity as the near-term ceiling. See the sickle cell and SMA briefs, or commission a model scoped to your asset.