Roche and Ionis reported on September 22 that the prespecified interim analysis of the Phase 3 IMAgINATION trial met its primary endpoint. In 459 adults with primary IgA nephropathy at high risk of progression, sefaxersen produced a statistically significant and clinically meaningful reduction in proteinuria against placebo at 37 weeks, with no new safety signals. Proteinuria was measured as the 24-hour urine protein-to-creatinine ratio.
The announcement did not include the size of that reduction.
In most diseases a missing effect size is a detail that arrives later with the conference abstract. In IgA nephropathy it is the substance of the claim, because proteinuria reduction is the measurement the entire field is currently built on, and the companies competing in it have published their numbers.
The number is the whole argument in this disease
The FDA has accepted proteinuria as a surrogate endpoint for accelerated approval in IgA nephropathy, on the condition that the sponsor later shows the drug preserves kidney function. That trade is why the reductions matter: they are the basis on which four therapies have been cleared or are close to it, and each one arrived with a figure attached.
Otsuka's Voyxact, an APRIL inhibitor better known as sibeprenlimab, was approved in November 2025 on a 51 percent placebo-adjusted reduction in proteinuria at nine months from the VISIONARY trial, given as a 400 milligram subcutaneous injection every four weeks. Vertex's povetacicept produced a 49.8 percent placebo-adjusted reduction at week 36. Vera Therapeutics' atacicept, which we analyzed when its confirmatory data landed on September 15, delivered a 41.8 percent between-group reduction at week 36. Novartis' Fabhalta, an oral factor B inhibitor, reduced proteinuria by 38.3 percent at nine months.
That is the scale any new entrant is placed on, and the reporting on the interim result noted the arithmetic problem: Roche said the drug has best-in-class potential, and did not release the number that would test the claim. A reduction of 25 percent and a reduction of 45 percent both satisfy the sentence the company published. The difference decides whether sefaxersen is a competitor to the incumbents or a later-line option behind them.
What an antisense drug does differently
The mechanism is the most distinctive thing about the program, and it is one competitors cannot copy by reformulating. Fabhalta binds complement factor B directly. Sefaxersen, formerly IONIS-FB-LRx, is a liver-directed antisense oligonucleotide that stops the liver from producing the protein in the first place, by attaching to the messenger RNA that encodes it. Blocking production and blocking the finished protein arrive at the same pathway from opposite ends, and factor B drives the alternative complement cascade that damages kidney tissue in this disease.
The delivery differs too. Sefaxersen is a once-monthly subcutaneous injection designed for self-administration, with loading doses on days 1, 15 and 29 before the monthly schedule begins. Fabhalta is an oral capsule taken twice a day. For a chronic kidney disease where patients already manage multiple prescriptions, a monthly autoinjector is a convenience argument, and it is the one Roche has been making since before the data existed. The counterweight is that Fabhalta can raise cholesterol, and Fierce Biotech reported that sefaxersen showed no observed effect on cholesterol in its Phase 2 data, which would let a prescriber separate the two factor B drugs on tolerability rather than only on convenience.
The program has a longer history than the current race suggests. GSK held rights to the asset and walked away in 2017. Roche paid $75 million for an option in 2018 and $55 million upfront to license it for IgA nephropathy in 2022, and Ionis stands to receive up to $400 million in milestones plus royalties if the drug reaches the market. BioPharma Dive noted that the arrangement came out of a broader Roche-Ionis partnership spanning Huntington's and Alzheimer's programs.
The confirmatory readout is the part still running
IMAgINATION stays blinded to week 105, where the key secondary endpoint is the change in estimated glomerular filtration rate, the measure of how well the kidneys are filtering. The registry lists the estimated primary completion as August 31, 2026 and the estimated study completion as March 2029, with participants able to move to open-label treatment after week 105 or a common close, whichever arrives first.
That schedule is the same one every drug in this class is on. An accelerated approval granted on week-37 proteinuria still requires the two-year data, and the label that comes with the first approval carries a sentence saying it has not been established whether the drug slows long-term kidney function decline. Vera Therapeutics spent September removing that sentence from its own label with final ORIGIN 3 data. Sefaxersen is at the front of the same path, which means an approval based on this interim would arrive with the same conditional language attached until the eGFR slope reports.
Roche's own framing of the path has been consistent. Jay Garg, the company's global head of nephrology, rheumatology and cardiovascular, said at a March investor event that the week-37 proteinuria data will "hopefully lead to accelerated approval," and the company said the interim results will go to an upcoming medical congress and to health authorities.
A class that has already set its own bar
IgA nephropathy has gone from a disease with no approved targeted therapy to one with several in under two years. Fabhalta received accelerated approval in 2024 and was converted to full authorization this year. Voyxact followed in November 2025, atacicept in July 2026, and Vertex's povetacicept is expected to be the next decision. The mechanisms do not overlap neatly: APRIL inhibition, dual BAFF and APRIL inhibition, oral factor B inhibition, and now an antisense approach to factor B production.
The prices reflect the scarcity the approvals are relieving. BioPharma Dive put recently launched IgA nephropathy therapies at roughly $400,000 to $500,000 per year. In a market with several drugs that all reduce proteinuria by a substantial margin, the differentiators narrow to magnitude, dosing, tolerability and payer access, and the first of those is the one Roche has not disclosed.
Analysts have been sizing the opportunity without the data. Roche's head of investor relations, Bruno Eschli, said at a March company event that sefaxersen appears in most analyst models with sales of 400 million euros forecast by 2030, while Roche's internal view runs to peak sales of 1 billion to 2 billion euros, figures set before the readout. Myles Minter of William Blair told BioPharma Dive that monthly autoinjector dosing is "relatively competitive," noting that only Vertex's candidate offers a similar profile, and Trung Huynh of RBC Capital Markets grouped sefaxersen among Roche pipeline assets he considers underappreciated by investors.
What the congress abstract will have to answer
The word that carries weight in the announcement is prespecified. The interim analysis plan, including the week-37 timing and the endpoint, was fixed in advance, which protects the result against the criticism that the goalposts moved. It does not protect the comparison against the other limitation in this field: none of these drugs has been tested head to head against another. The 38 to 51 percent band is assembled from separate trials with different enrollment criteria, different baseline proteinuria levels and different measurement windows, so even a published number would come with the caveat that cross-trial comparisons are approximations rather than rankings.
What the congress presentation will settle is narrower and still important. If it includes the percentage reduction and the confidence interval, the market can place sefaxersen inside the band or outside it. If it repeats the press release's language without figures, the field will be comparing an unknown quantity to four known ones for another quarter at least, while the trial continues toward the eGFR readout that determines whether any of these drugs keeps its approval.
The kidney function data is where the class's real differentiation will appear, and it is also where the timeline is longest. Proteinuria at 37 weeks is a proxy; a slope measured over 105 weeks is the outcome patients care about, because it is the difference between a drug that lowers a lab value and one that delays dialysis. The field has approved four therapies on the promise that the two are linked, and the confirmatory trials are the test of the promise itself.
The hesitation is a competitive choice
There is a version of this announcement in which Roche simply had not finished the analysis and released what it had. There is another in which a company with the second factor B drug on the market chose to lead with the endpoint rather than the effect size, because the endpoint is unambiguous and the effect size would invite a direct comparison with an oral pill that already has full approval in the same indication.
Either way, the decision has a cost that is measurable. Investors who have been told the drug is in most models at 400 million euros and could reach 2 billion have no way to test those figures against a competitor set whose numbers are public. Clinicians deciding where a monthly injection fits have the same gap. And patients, who are the actual audience for a best-in-class claim, have no basis to weigh a new option against the three they can already be prescribed.
The company's position is that the data will be presented at a medical congress and shared with regulators, which is the ordinary path and the one that resolves the question. Until then, the accurate description of the IMAgINATION interim is an endpoint met, on a surrogate that the field has agreed to accept, with the measurement that would place it among its peers still sealed inside a blinded trial.
That contrast is the story of IgA nephropathy in 2026. Trutakna's confirmatory trial delivered the number its label was missing, and the drug became the first in the class with two-year kidney function data behind it. Sefaxersen arrives at the same door from the other side: an endpoint met, a filing path opened, and the number that would tell patients and prescribers what they are choosing between withheld until a room of specialists hears it first.
Primary sources
- Genentech, Positive interim data shows Genentech's sefaxersen significantly reduces proteinuria in people with IgA nephropathy, for the interim result, the trial size and the plan for the two-year eGFR readout.
- European Biotechnology, Roche hits Phase 3 IgAN endpoint but holds back data, for the withheld effect size and the competitor proteinuria reductions.
- Fierce Biotech, Roche's phase 3 win tees up fight with Novartis for blockbuster kidney disease market, for the factor B comparison with Fabhalta and Roche's commercial expectations.
- BioPharma Dive, Roche, Ionis RNA drug scores in kidney disease study, for the Ionis milestone terms, the analyst commentary and the pricing context.
- ClinicalTrials.gov, IMAgINATION: sefaxersen in primary IgA nephropathy (NCT05797610), for the trial design, dosing schedule, endpoints and estimated dates.
- U.S. Food and Drug Administration, Drug Trials Snapshot: Voyxact, for the sibeprenlimab accelerated approval basis.