Uthando Nethemba

CNS antisense oligonucleotides — tofersen missed its clinical primary yet was approved, and eplontersen missed its and was not: same modality, opposite regulatory outcome

—

by

itembalabantu
in

Evidence-first notes on bioscience and deep tech, at the edge of the lab and the market. Information only — not investment advice, not medical advice. All trial results, reduction figures and safety events are attributed to the sponsoring trial, company or regulator; cross-trial and cross-indication comparisons are not head-to-head. Surrogate-biomarker results (neurofilament light) are kept separate from clinical endpoints (ALSFRS-R, survival).

The 30-second version

  • What. Tofersen (Qalsody, Ionis/Biogen) is an intrathecal antisense oligonucleotide (ASO) that knocks down mutant SOD1 in SOD1-linked ALS. Its pivotal trial VALOR MISSED its clinical primary endpoint — the 28-week ALSFRS-R difference was 1.2 points (95% CI −3.2 to 5.5), P=0.97 (Miller et al., NEJM 2022;387:1099–1110). Despite that miss, the FDA granted accelerated approval on 2023-04-25 on the plasma neurofilament-light (NfL) surrogate, not on a clinical endpoint. (It did lower the causal protein — CSF total SOD1 fell to a tofersen/placebo ratio of 0.62.)
  • So what. The central thread: tofersen (ASO, CNS — primary MISS, then APPROVED on a surrogate) versus eplontersen in ATTR cardiomyopathy (ASO, heart — primary MISS, then NOT approved for CM). Same modality, opposite regulatory outcome. The divergence is NOT a verdict on the modality. It decomposes into three drivers: (a) unmet need/rarity — no approved disease-modifying therapy in SOD1-ALS versus tafamidis-established standard of care in ATTR-CM; (b) surrogate acceptance — NfL was accepted for CNS, while no cardiac surrogate is accepted for approval (patisiran met a 6-minute-walk endpoint yet received a CRL); (c) regulatory pathway — accelerated-on-surrogate with a confirmatory trial, versus a hard-outcome confirmatory trial with no surrogate fallback.
  • Now what. The accelerated approval is conditional: clinical benefit is NOT confirmed. The confirmatory trial is ATLAS, testing whether NfL-guided early intervention in presymptomatic SOD1 carriers delays or prevents clinical onset — that is the real test of the surrogate logic. Read tofersen honestly: cause-protein and NfL knockdown are demonstrated, but clinical benefit is unproven. “CNS ASO is superior to cardiac ASO” is an unsupported misread of these two divergent regulatory outcomes.

The five-minute read

The sharpest symmetry in the series — same modality, opposite outcome

An earlier part of this series examined a target-matched, knockdown-depth-matched pair (siRNA versus ASO against TTR) whose outcomes diverged. This part goes one step further: two drugs of the same modality (ASO) that both missed the same kind of endpoint (the clinical primary), yet whose regulatory outcomes went in opposite directions. Tofersen (Qalsody, Ionis/Biogen, SOD1-ALS) missed its VALOR primary (ALSFRS-R, P=0.97) and was nonetheless granted FDA accelerated approval (2023-04-25) on the plasma NfL surrogate. Eplontersen (AstraZeneca/Ionis, ATTR-CM) missed the CARDIO-TTRansform primary (CV death plus CV events) and was not approved for the cardiomyopathy indication (full data awaited).

The surface headline reads “CNS ASO approved, cardiac ASO not.” The thesis of this part is that what diverged was neither the modality nor the fact of missing a primary, but three things: unmet medical need and disease rarity; the regulatory acceptability of the NfL surrogate; and an accelerated-approval-on-surrogate pathway versus a hard-outcome confirmatory pathway. Decomposing that divergence precisely is the core of this part — it is the regulatory-layer instance of the series’ central question: does a surrogate success translate into clinical benefit?

The no-fabrication alarm (the most common error)

VALOR missed its clinical primary endpoint. The statement “tofersen succeeded in its trial and was approved” is wrong. The approval is an accelerated approval based on the plasma NfL surrogate, and clinical benefit is not confirmed (the confirmatory trial ATLAS is ongoing). This refuted proposition is surfaced at the Tier-1/headline level, not buried.

Axis tofersen (SOD1-ALS) eplontersen (ATTR-CM)
Target / modality mutant SOD1 knockdown, ASO (RNase H), intrathecal TTR knockdown, ASO (RNase H), subcutaneous
Primary endpoint result ALSFRS-R MISSED (P=0.97) CV death + CV events MISSED (press release)
Unmet medical need ~2% of ALS, no approved DMT, uniformly fatal tafamidis (stabilizer) already standard of care
Surrogate acceptance NfL accepted as basis for accelerated approval No cardiac surrogate accepted for approval (even 6MWT → patisiran CRL)
Regulatory pathway Accelerated approval (surrogate + confirmatory ATLAS) The hard-outcome confirmatory trial itself is the primary — no fallback on a miss
Outcome Accelerated approval (2023-04) Not approved for CM (full data awaited)
Same modality (both ASO), same “primary missed” fact — opposite regulatory outcome. The divergence is driven by unmet need, surrogate acceptance and pathway design, not by the modality. This is a cross-trial, cross-indication contrast — not a head-to-head comparison. “CNS ASO superior / cardiac ASO inferior” is an unsupported misread. NfL reduction is pathway engagement, not confirmed clinical benefit (ATLAS pending).

Deep dive

1. Background — tofersen and VALOR (the primary MISS, the core case)

Trial: VALOR (NCT02623699), N=108 randomized, with the primary analysis population being the faster-progression subgroup, N=60 (tofersen 39 / placebo 21). Tofersen 100 mg was given intrathecally. The primary endpoint was the change in ALSFRS-R total score from baseline to week 28 in the faster-progression subgroup (Miller et al., N Engl J Med 2022;387:1099–1110, 2022-09-22, DOI 10.1056/NEJMoa2204705).

  • Primary endpoint (as reported): MISSED. The adjusted mean ALSFRS-R change was −6.98 (tofersen) versus −8.14 (placebo); adjusted mean difference 1.2 points (95% CI −3.2 to 5.5), P=0.97 (joint rank test with multiple imputation). The direction slightly favored tofersen but was not statistically significant. Per the paper, “statistical significance was not achieved for the primary end point,” so all subsequent secondary endpoints are treated as non-significant under the testing hierarchy.
  • Target-engagement surrogates (secondary, nominal): CSF total SOD1 — tofersen/placebo adjusted geometric mean ratio 0.62 (95% CI 0.49–0.78), i.e. the toxic SOD1 protein was actually lowered (cause-protein knockdown confirmed). Plasma NfL — a large decrease on tofersen and an increase on placebo (reported at roughly −60% versus +20%; the exact adjusted ratio is unverified, §9).
  • Open-label extension (52-week pooled analysis): the early-start group showed numerically less decline than the delayed-start group (which began 28 weeks later) on ALSFRS-R, percent-predicted SVC and handheld-dynamometry megascore. This is hypothesis-generating (not randomized or blinded — suggestive of an early-intervention benefit, not confirmation).

Interpretation: VALOR established the surrogate (it lowered the causal protein SOD1 and the axonal-injury marker NfL) but did not show a significant benefit on the 28-week clinical endpoint (ALSFRS-R). This is the ALS version of the series’ central question — structurally identical to the ATTR-CM situation where “TTR was lowered but the hard outcome diverged.”

2. What this part newly establishes — tofersen’s accelerated approval and the NfL surrogate logic (the regulatory core)

The FDA granted accelerated approval to Qalsody (tofersen) for SOD1-ALS on 2023-04-25. The basis was not a clinical endpoint but the reduction in plasma NfL (neurofilament light) — accepted as a surrogate “reasonably likely to predict clinical benefit” (Biogen press release, 2023-04-25).

  • Confirmatory-trial requirement: accelerated approval is conditional on a confirmatory trial. Tofersen’s confirmatory axis is ATLAS — enrolling presymptomatic SOD1 mutation carriers and intervening around the point of NfL rise, to test whether it delays or prevents clinical onset. Whether NfL-guided early intervention changes clinical outcomes is the real test of this surrogate logic.
  • Why NfL was accepted (four grounds):
    1. Causal proximity: SOD1-ALS is a monogenic cause — mutant SOD1 is a validated causal target. Because tofersen directly lowers that causal protein (CSF SOD1), NfL (axonal injury) sits downstream on the causal path.
    2. Unmet medical need: the SOD1 form is ~2% of all ALS, uniformly fatal, with no approved DMT — the textbook condition for regulatory flexibility.
    3. Biological plausibility: NfL is a quantitative marker of neuroaxonal degeneration, and tofersen lowered it substantially and consistently.
    4. OLE directional consistency: the early-start group’s numerical advantage is not inconsistent with the surrogate logic (though not confirmatory).

The NfL surrogate debate: NfL is a disease-nonspecific marker of axonal injury. Lowering NfL is pathway engagement, not confirmation of clinical benefit — the fact that the VALOR clinical primary was missed is exactly what shows that uncertainty. The accelerated approval is a conditional approval that explicitly states clinical benefit is not yet proven.

3. The central thread — decomposing the symmetry: tofersen (approved) versus eplontersen (not approved)

Both drugs are ASOs and both missed their clinical primary. Yet the regulatory outcomes are opposite. The divergence decomposes along three axes (this is a cross-trial, cross-indication contrast — not head-to-head, §6/§9).

(a) Unmet medical need and disease rarity — the primary driver of regulatory flexibility. SOD1-ALS is a fatal ultra-rare disease with no approved DMT, so the FDA was flexible toward a surrogate-based accelerated approval. ATTR-CM already has tafamidis (a stabilizer) established as standard of care, so a new drug must prove add-on benefit on top of it (the “add-on threshold”). The size of the unmet need governs the regulatory bar.

(b) Surrogate acceptance — NfL accepted, cardiac surrogate not. This is the decisive asymmetry. NfL was the explicit basis for the SOD1-ALS accelerated approval. For the heart, there is essentially no surrogate accepted as a basis for approval — patisiran (APOLLO-B) met a surrogate/functional endpoint (6-minute-walk median difference 14.69 m, P=0.02) yet still received a CRL in 2023-10 (cited reasons: clinical meaningfulness, short follow-up). Cardiac programs are therefore required to show hard outcomes (CV death, hospitalization), and both ASO and siRNA are tested at that bar.

(c) Regulatory pathway — accelerated approval versus hard-outcome confirmatory trial. Tofersen took the accelerated-approval pathway (approval on a surrogate, with a confirmatory-trial obligation): even though a small (N=108), short (28-week) trial failed to hit its clinical primary, the structure is approve on NfL, then confirm via ATLAS. Eplontersen’s CARDIO-TTRansform (N=1,432, ~2.7 years) was itself designed as a hard-outcome confirmatory trial — missing the hard outcome leaves no fallback onto a surrogate. For the same “primary missed,” the pathway design determined the outcome.

Sub-conclusion: the cause of the symmetry is not the modality (both are ASOs). It is the three-way difference in unmet need, surrogate acceptance and regulatory pathway. Therefore “CNS ASO superior / cardiac ASO inferior” is an unsupported misread. (Indeed eplontersen is an approved agent in the ATTR polyneuropathy indication via NEURO-TTRansform — §9 unverified — evidence that a different indication carries a different bar.)

4. The platform lineage — nusinersen (a success precedent) and tominersen (a caution case)

Intrathecal ASO became a validated modality for crossing the CNS delivery barrier thanks to two events before tofersen.

Nusinersen (Spinraza, Ionis/Biogen, SMA) — the success precedent that validated the platform. An intrathecal ASO that corrects SMN2 splicing to raise SMN protein. ENDEAR (NCT02193074, Phase 3, 2:1 randomized, sham-controlled): HINE motor-milestone responders 51% versus 0% (P<0.0001), event-free survival HR 0.530 (P=0.0046), overall survival HR 0.372; stopped early for efficacy (Finkel et al., N Engl J Med 2017;377:1723–1732). This proved that intrathecal ASO can deliver clear clinical benefit (motor milestones, survival) — the foundation of platform confidence. (But SMA differs from tofersen in target and mechanism, so the result cannot be transplanted — nusinersen won on a hard clinical benefit, tofersen on a surrogate.)

Tominersen (Roche/Ionis, Huntington’s disease, HTT) — the caution case for a knockdown modality. An intrathecal ASO that lowers both normal and mutant huntingtin (HTT). GENERATION-HD1 (Phase 3) was halted on 2021-03-22 on IDMC recommendation on a benefit/risk basis (no new safety signal). Post-hoc analysis suggested younger, less-advanced, lower-CAG-repeat patients might benefit at lower dose/frequency (interpreted alongside a worsening trend in the high-dose arm). It was redesigned as GENERATION-HD2 (Phase 2, ~301 patients, 60/100 mg q16w, focused on early patients; the IDMC later stopped the 60 mg arm and continued 100 mg; expected to complete in 2026, final efficacy undisclosed).

The lesson of the three drugs: intrathecal ASO is validated as a modality (nusinersen), but outcomes diverge dramatically by target and disease — SMA (hard benefit), SOD1-ALS (surrogate-based approval), HTT (halted, redesigned). Modality maturity (TRL) and the translational success of an individual target are separate things — a restatement of the series’ central proposition.

5. Commercialization and competitive context (TRL, related companies)

  • Maturity (TRL frame): intrathecal ASO is a mature modality (nusinersen carries hard clinical benefit; roughly TRL 8–9 as a delivery platform). Tofersen’s cause-protein and NfL knockdown are demonstrated, but clinical benefit is unconfirmed, so the SOD1-ALS clinical-benefit maturity sits earlier (confirmatory ATLAS pending). The gating layer is the surrogate-to-clinical translation, not the delivery.
  • Ionis (IONS) / Biogen (BIIB): tofersen (Qalsody) accelerated approval (2023-04-25 on NfL, confirmatory ATLAS) and nusinersen (Spinraza) are the flagship intrathecal-ASO franchises.
  • Roche (with Ionis): tominersen (HTT) — GENERATION-HD1 halted (2021-03-22) and redesigned as GENERATION-HD2 (Phase 2, ongoing to 2026).
  • AstraZeneca (AZN) / Ionis: eplontersen — CARDIO-TTRansform CM primary missed (press release; full data awaited). The polyneuropathy indication is a separately approved lineage (§9 unverified).
  • Company statements are limited to neutral, trial-attributed description; regulatory-outcome and competitive statements are not buy/sell signals. Deal terms and exact figures are attributed to the trial, company or regulator.

6. The skeptic’s bottom line (Skeptic block)

  • VALOR missed the clinical primary (P=0.97): the approval is an NfL-surrogate accelerated approval, not confirmation of clinical benefit. Do not say VALOR succeeded.
  • NfL is a nonspecific pathway marker: lowering it is not clinical benefit; it remains unconfirmed until ATLAS reads out.
  • tofersen versus eplontersen is cross-trial, cross-indication: indication, standard of care and trial design differ — do not assert a modality verdict. “CNS ASO superior” is unsupported.
  • The OLE early-start advantage is unblinded and hypothesis-generating, not confirmatory.
  • Tominersen was halted and redesigned (a high-dose worsening trend) — a target-specific risk of the knockdown modality.
  • Neutral-framing note: “CNS ASO superior / cardiac ASO inferior” security implications, and any statement asserting confirmed clinical benefit from the NfL accelerated approval, are blocked as red-zone. Cross-trial ≠ head-to-head; NfL surrogate is kept separate from the clinical endpoints (ALSFRS-R, survival).

7. What to watch (falsifiable)

  • P1 — ATLAS (presymptomatic SOD1) readout: if NfL-guided early intervention significantly delays or prevents clinical onset, the NfL surrogate logic is confirmed — a neurological success case for “target engagement translates to clinical benefit.” If it does not affect onset, the position moves to “NfL is a pathway marker but insufficient as a clinical surrogate” (surrogate untranslated).
  • P2 — eplontersen full data (ATTR-CM, full dataset): if benefit reappears in the monotherapy stratum, the cause of the symmetry is reinforced as the add-on threshold / SoC (not efficacy) — strengthening the “not a modality” thesis. If it also vanishes in monotherapy, the interpretation shifts toward a cardiac-target-specific factor.
  • P3 — GENERATION-HD2 final efficacy (2026): if clinical benefit appears at 100 mg in early patients, the knockdown modality moves to being a function of target, dose and patient selection (HTT can work too). If it misses again, this strengthens the view that HTT has a larger translational bottleneck than SOD1 or SMN2 (target specificity).

References

Disclosure

This post is for information only and is not investment advice, and not medical advice. Treatment decisions should always be made with your own clinician.

COI note: this post describes listed companies and their sponsored trials — Ionis (IONS), Biogen (BIIB), Roche, AstraZeneca (AZN) — in a descriptive, neutral context. Every trial result, reduction figure and safety event is attributed to its source: VALOR and OLE figures to Miller et al. (NEJM 2022); ENDEAR to Finkel et al. (NEJM 2017); the tominersen halt to the Roche press release (2021-03-22); the tofersen (Qalsody) accelerated approval to the FDA / Biogen press release (2023-04-25); the eplontersen ATTR-CM result to the AstraZeneca press release (full data awaited). The tofersen-versus-eplontersen contrast is a scientific comparison of trial design, indication, unmet medical need and surrogate acceptability — it is not an assertion of modality superiority. Cross-trial and cross-indication comparisons are not head-to-head. The NfL accelerated approval does not confirm clinical benefit (ATLAS ongoing). Quantitative claims are attributed to the trial, author or press release. Competitive and regulatory-outcome statements are factual, neutral descriptions and are not buy/sell implications for any security. The author holds no position in, and has no financial interest in, the companies named.

Comments

Leave a comment