brain/
sourcestock-market

Autoresearch: eplontersen ATTR-CM / CARDIO-TTRansform ESC 2026 + nucresiran + ALNY guidance

ESC 28–30 Aug 2026 published CARDIO-TTRansform: overall miss; stabilizer (all tafamidis) subgroup remains a clinical and NT-proBNP null. No primary isolates ASO toxicity. ALNY −28–30% is AMVUTTRA 2L guidance, not a silencer readout. TRITON-CM still recruiting / adjunctive / no outcomes data.

Source

Autoresearch: eplontersen ATTR-CM / CARDIO-TTRansform ESC 2026 + nucresiran + ALNY guidance

Generated by /autoresearch on 2026-09-19. Synthesized across 3 rounds from 13 web pages (plus 4 fetch failures), Grokipedia skipped per method (no encyclopedia-as-source). See Provenance. Treat as raw material — review before promoting. Do not re-rate ALNY/BBIO/PFE/AZN/IONS. Context: vault/projects/stock-market (P1 open step 3 + step 2 flag on eplontersen-failure-to-alnylam-silencer-consolidation)

Summary

Full CARDIO-TTRansform data were presented at ESC Congress 2026 in Munich on 28 August (Hot Line) and 30 August (late-breaking stabilizer analysis) and simultaneously published in the New England Journal of Medicine and Nature Medicine. The overall primary endpoint remained a miss (rate ratio 0.89; 95% CI 0.73–1.09; P=0.277) despite TTR suppression. Baseline stabilizer use — all tafamidis — modified the result: monotherapy nominally positive (RR 0.71; P=0.012); on-stabilizer no added benefit (RR 1.14; P=0.39), and NT-proBNP on that background was similar to placebo (P=0.08). So the July “total null on tafamidis, including CV biomarkers” read holds for the tafamidis subgroup and does not hold for monotherapy or the overall Ionis topline biomarker claim.

No AZN, IONS, ESC, NEJM, or Nature Medicine primary attributes the miss to ASO-specific toxicity. Safety was described as favorable / consistent with prior eplontersen use. Sponsors and the PI attribute the miss to contemporary stabilizer background (and, in the secondary paper, pathway redundancy). Alnylam, on its 30 July 2026 Q2 call, attributes the miss to “molecule and study-specific issues” (knockdown depth/variability, population, design) and contrasts RNAi with ASO as silencing modalities, not as a documented cardiotoxicity finding. That is not a primary isolation of ASO toxicity as the cause of the fail.

The Biotech Hangout “shares dropping 30% following disappointing updated guidance” flag is the 30 July 2026 Q2 print, not eplontersen: TTR product-revenue guide cut $4.4–4.7B → $4.2–4.5B because U.S. AMVUTTRA second-line demand normalized after 2025 pent-up switch volume. Cause is commercial launch-curve, not a silencer-class clinical readout. Adjacent, not load-bearing, to the adjunctive-silencer thesis.

Nucresiran’s ATTR-CM CVOT (TRITON-CM, NCT07052903) is still recruiting, still allows tafamidis/acoramidis, still has no outcomes readout (primary completion ~May 2030). Target N was expanded in Q1 2026 (pre-eplontersen miss) from 1,250 to ~1,750 via a pre-specified option. No confirmed post-ESC protocol rewrite as of 19 September 2026; management said it might adapt after reviewing full eplontersen data and, at Morgan Stanley mid-September, that enrollment adjustments might be communicated by year-end.

Honesty: Step 3 stays open. No primary isolates ASO-specific toxicity as the reason eplontersen failed on tafamidis. Nucresiran has not read out. The ESC stabilizer-modification + HELIOS-B tafamidis-subgroup contrast sharpens the question; it does not close it.

Findings

ESC published the full CARDIO-TTRansform package (28–30 Aug 2026)

The dated catalyst is no longer pending. ESC’s own press release, dated 28 August 2026 in Munich, states the trial “did not meet its primary efficacy endpoint” in a Hot Line session that day: 381 primary events in 210 eplontersen patients vs 392 events in 231 placebo patients (rate ratio 0.89; 95% CI 0.73–1.09; P=0.277), despite circulating TTR suppression, in 1,432 patients at 130 centres in 20 countries (ESC press release, 28 Aug 2026). ACC’s same-day recap adds simultaneous NEJM publication (ACC, 28 Aug 2026). Healio cites the NEJM paper as Fontana et al., doi:10.1056/NEJMoa2608510 (Healio, 28 Aug 2026). A second ESC late-breaking talk (García-Pavía, 30 Aug) on stabilizer subgroups was published as García-Pavía et al. in Nature Medicine, doi:10.1038/s41591-026-04670-6 (Nature Medicine AAP; ARCI recap, 2 Sep 2026).

This is the dataset AZN/IONS promised on 9 July 2026 when they reported the topline miss and said full results would go to ESC in August (AstraZeneca, 9 Jul 2026; Ionis exhibit 99.1 / SEC, 9 Jul 2026).

On tafamidis background the clinical null stands — including NT-proBNP

Fifty-seven percent of randomized/treated patients (819/1,432) were on a TTR stabilizer at baseline; all of those baseline stabilizers were tafamidis. Another ~24% per arm started a stabilizer during the trial (AZN: 24% each arm; Nature Medicine: 174/305 and 174/308 of the no-baseline-stabilizer patients, ~57% drop-in, “predominantly tafamidis”) (AstraZeneca, 9 Jul 2026; García-Pavía et al., Nature Medicine).

Prespecified interaction by baseline stabilizer: P_interaction = 0.017 on the primary composite.

  • No baseline stabilizer (n=613): 172 vs 221 primary events; RR 0.71 (0.54–0.93); P=0.012. 6MWT LSMD +29.9 m; KCCQ-OS LSMD +7.6; NT-proBNP increase mitigated vs placebo from week 49 (P<0.001 for change) (García-Pavía et al.).
  • Baseline stabilizer / all tafamidis (n=819): 209 vs 171 primary events; RR 1.14 (0.85–1.53); P=0.39. 6MWT LSMD +12.0 m (P=0.07); KCCQ-OS LSMD +1.1 (P=0.37). NT-proBNP “similar” through week 140 (P=0.08) (García-Pavía et al.; Healio quotes Maurer RR 1.14 vs 0.71, P_interaction=0.017).

Maurer at ESC: “Primary results appeared to be influenced by baseline stabiliser use with no clear benefit in patients receiving background stabiliser therapy” (ESC; ACC). At the press conference: “combining the silencer eplontersen with a stabilizer did not result in incremental benefit in terms of morbidity and mortality” (Healio).

TTR knockdown worked on both backgrounds. End-of-trial median absolute TTR on eplontersen was 6.5 mg/dL (no baseline stabilizer) vs 7.1 mg/dL (on stabilizer); on-treatment 6.3 vs 6.9. The secondary paper says differences in TTR reduction are “unlikely to account for” the subgroup split (García-Pavía et al.). ARCI’s patient-facing recap of the ESC talks puts overall TTR lowering at “about 79%”; Healio reports mean change −69.7% at 140 weeks (ARCI; Healio).

July 9 vs July 10 tension, now resolved at the subgroup level. Ionis/AZN topline said that in the overall population “multiple secondary, imaging and biomarker analyses favored eplontersen” (Ionis SEC exhibit). The 10 July Hangout “no benefit on CV biomarkers” line matches the tafamidis-background NT-proBNP result, not the monotherapy or overall claim. ESC did not convert the on-tafamidis arm into a biomarker win.

Acoramidis cannot be read from this trial: Healio notes tafamidis “was the most prevalent stabilizer” and conclusions cannot be drawn about Attruby (Healio). Nature Medicine is stronger: baseline stabilizer use was all tafamidis.

Sponsors and investigators blame stabilizer background / pathway overlap — not ASO toxicity

AZN (Sharon Barr): the trial was “designed to examine the role of Wainua … on top of today’s standard of care”; it “did not meet its primary objective” (AstraZeneca, 9 Jul 2026). Ionis CEO Brett Monia: “these findings reflect the rapidly evolving treatment landscape, in which contemporary ATTR-CM patients are widely treated with stabilizers” (Ionis SEC exhibit). PI Maurer on 9 July: the data “provide important clarity for the field”; at ESC he names baseline stabilizer use as the influence on the primary (Ionis SEC exhibit; ESC).

Nature Medicine discussion (AZN + Ionis authors including Jersey Chen and Sotirios Tsimikas): silencers and stabilizers “act at different places within the same biological pathway … likely accounting for absence of additional benefit of eplontersen on top of background stabilizer therapy.” They explicitly reject knockdown failure as the subgroup explanation. They contrast HELIOS-B’s capped 40% baseline stabilizer use and lower drop-in vs CARDIO-TTRansform’s unrestricted contemporary population. HELIOS-B’s own stabilizer-subgroup RR for a similar composite was 0.78 (0.50–1.22; P=0.28; n=259) — a non-significant point estimate they say CARDIO-TTRansform’s larger n=819 now makes “not evident” for eplontersen (García-Pavía et al.).

Safety language is the opposite of a toxicity-cause story. ESC: “generally well tolerated, with a safety profile consistent with previous results” (ESC). Nature Medicine: “safe and well tolerated … similar rates of serious TEAEs and TEAEs relating to study drug discontinuation” regardless of baseline stabilizer. Healio: SAE 57.8% eplontersen vs 59.4% placebo; Medscape (secondary recap of the Hot Line) has discontinuations 6.4% vs 7.4% (Healio). ARCI: “no new safety concerns … whether or not someone was also on a stabilizer” (ARCI).

No fetched AZN, IONS, ESC, NEJM, or NatMed sentence attributes the fail to ASO-specific cardiotoxicity. The July Maraganore ASO-class framing remains a podcast inference, not a trial-primary finding.

Alnylam’s competing attribution is molecule / knockdown / design — still not a toxicity isolation

On the 30 July 2026 Q2 call (three weeks after the eplontersen topline, four weeks before ESC), CEO Yvonne Greenstreet: CARDIO-TTRansform “does not alter our conviction in the TRITON-CM study”; they will “carefully review the full eplontersen data set … and adapt our study plan, if appropriate” (Motley Fool transcript of ALNY Q2, 30 Jul 2026). CSO Pushkal Garg: “likely attributable to a combination of molecule and study-specific issues”; “RNAi therapeutics are fundamentally different than antisense oligonucleotides”; “we’ve also seen that the safety profiles of these two approaches differ as well”; they would look at ESC for “safety … given what we know about ASOs in the past” and for knockdown depth/variability, NYHA mix, and endpoint components. He modeled ~67% of eplontersen-PN patients vs ~82–84% of vutrisiran patients reaching ≥80% TTR knockdown, and >99% for nucresiran at 95% median knockdown. Bottom line on that call: “we don’t believe that the top line results … negate the hypothesis … of using a silencer for ATTR-CM patients who are already on a stabilizer. More likely … the type and depth of silencing, along with aspects of the study design” (same transcript).

That is sponsor-competitor advocacy, not an independent isolation of ASO toxicity. After ESC, Alnylam did not publish a “we found ASO tox” note. Its 30 August ESC release is a HELIOS-B tafamidis-subgroup late-breaker: among 654 treated patients, 259 (40%) on tafamidis at baseline; “treatment effect … consistent irrespective of baseline tafamidis use”; HELIOS-B “was not powered to establish the benefit of vutrisiran specifically in the population of patients receiving background tafamidis”; combination strategies “warrant further evaluation” (Alnylam ESC PR, 30 Aug 2026). That is a directional RNAi-on-tafamidis claim with an explicit underpower caveat — it does not close the adjunctive-silencer question, and Nature Medicine already flagged the HELIOS-B stabilizer CI as wide.

ALNY −28–30% after 30 July guidance is a 2L AMVUTTRA commercial cut, not a silencer clinical event

Primary: Alnylam Q2 2026 release, 30 July 2026. TTR net product revenue guide $4,400–$4,700 million → $4,200–$4,500 million. Total net product revenues $4,900–$5,300 million → $4,700–$5,100 million. Collaboration/royalty guide raised $400–$500 million → $575–$625 million. Expense guide reiterated. Explicit cause: “updated outlook for AMVUTTRA in the second line segment of the U.S. market”; “growth in second line demand for AMVUTTRA moderated in early 2026 to what the Company now believes is a normalized level, following an early launch period that, with hindsight, benefited from pent-up demand from patients progressing on stabilizers who had been waiting for a new treatment option” (Alnylam Q2 PR; same table in SEC HTML earnings release).

CFO Jeff Poulton on the call: “Guiding the market’s expectations appropriately is important, and we didn’t get it right with our original guidance. We own that.” The $200 million midpoint cut is “the driver” of the TTR revision; first-line remains the commercial focus (~80% of new starts) (Motley Fool transcript). Q2 itself was not a collapse: TTR product revenue $1.03 billion (+89% YoY), first >$1B TTR quarter (Alnylam Q2 PR).

Secondary tape: RTT News same day had the stock at $204, down $82.62 / 28.83% from $286.62 close (RTT News, 30 Jul 2026); Zacks/Yahoo: −28.3% on the print plus the cut (Yahoo/Zacks). That is the 31 July Hangout show-note.

Orthogonal vs adjacent. The stated cause of the cut is exhausted 2025 2L pent-up demand, not CARDIO-TTRansform and not a nucresiran protocol change. Greenstreet even treated the eplontersen miss as competitive tailwind (“one less branded competitor”). So the −30% is not a clinical-silencer re-rating. It is a re-rating of AMVUTTRA’s second-line-on-stabilizer volume curve — the same commercial segment the adjunctive-null theoretically pressures. Do not collapse those two facts into one causal claim. Do not re-rate from this file.

Nucresiran TRITON-CM remains adjunctive, recruiting, unread-out

Official record: https://clinicaltrials.gov/study/NCT07052903 (direct WebFetch timed out; status corroborated by CenterWatch 22 Jul 2026 and HealTrials copy dated through late August). Phase 3, Alnylam-sponsored, status recruiting / active-recruiting. Start 2025-07-02; primary completion ~2026-listed as 2030-05-28; completion 2032-11-30. Primary: all-cause mortality + recurrent CV events (Andersen-Gill). Inclusion: “Patients may be receiving approved TTR stabilizers … (eg, tafamidis, acoramidis).” Exclusion: prior/current TTR-lowering therapy. Dose 300 mg SC q6M vs placebo. CenterWatch listing N=1,750 (CenterWatch NCT07052903; HealTrials copy of the CT.gov record).

Enrollment expansion was Q1 2026, before the eplontersen miss: Alnylam used a “pre-specified protocol option to expand target enrollment from 1,250 to approximately 1,750” because enrollment was faster than expected; still targeting a 2030 ATTR-CM launch if positive (Alnylam Q1 2026 PR). Garg on 30 July repeated the 1,750 figure and the event-driven design, and listed possible future adaptations (enrichment, analytic hierarchy) after ESC — “we may not need to do anything” (Q2 transcript). A 14 September 2026 Morgan Stanley conference summary still has only “possible enrollment adjustments to be communicated by year-end,” not a completed rewrite and no efficacy readout (Quartr summary). ARCI, writing 2 September after ESC, still lists TRITON-CM as an ongoing recruiter that should help the combo question (ARCI).

No nucresiran ATTR-CM outcomes data exist as of 19 September 2026.

Contradictions and open questions

  • Step 3 stays open. ESC/AZN/IONS/NatMed explain the miss as silencer-on-tafamidis / same-pathway redundancy. ALNY explains it as this ASO’s knockdown + this trial’s mix, and points to HELIOS-B/APOLLO-B combo signals. Neither camp published an ASO-toxicity causal claim. Nucresiran has not read out. Closing Step 3 would require a primary that isolates ASO toxicity or a TRITON-CM (or other siRNA) on-tafamidis outcomes result.
  • July “total biomarker null” is only true on tafamidis. Overall Ionis topline and the no-stabilizer NT-proBNP (P<0.001) contradict a global biomarker-null. The on-tafamidis NT-proBNP (P=0.08, “similar”) supports the original Matteis line for that subgroup.
  • HELIOS-B vs CARDIO-TTRansform on combination. ALNY ESC: vutrisiran effect “consistent” with/without tafamidis, underpowered for the combo slice (Alnylam, 30 Aug 2026). NatMed: HELIOS-B combo RR 0.78 (0.50–1.22), then CARDIO-TTRansform n=819 “not evident” for eplontersen. Cross-trial, not a head-to-head.
  • Knockdown as the ALNY escape hatch vs NatMed. Garg (pre-ESC) said insufficient knockdown is “one plausible contributor.” NatMed (post-ESC, with AZN/IONS authors) says end-of-trial absolute TTR on eplontersen was essentially the same with and without baseline stabilizer, so knockdown does not explain the within-trial split. It could still, in ALNY’s telling, explain why eplontersen combo failed while vutrisiran combo trended. Unresolved.
  • Post-ESC TRITON-CM adaptation is promised as a possibility (Q2 call; Morgan Stanley “by year-end”), not documented as a filed amendment in the sources fetched here. CT.gov HTML itself was not retrieved (timeout).
  • NEJM full HTML was not retrieved (timeout). Numbers above that depend on Fontana et al. come via ESC/ACC/Healio/ARCI, not a line-by-line read of the PDF.

Provenance

Rounds run: 3 of 3 (full)

Sub-questions by round:

Round 1 (broad survey):

  1. Was full eplontersen / CARDIO-TTRansform data presented at ESC August 2026, and what did the overall and tafamidis-background results show (including biomarkers)?
  2. What is TRITON-CM (nucresiran ATTR-CM CVOT) status: adjunctive/tafamidis allowed, any readout or protocol change?
  3. What caused the ~30% ALNY drop after “disappointing updated guidance” around 31 July 2026?
  4. Did any AZN/IONS/ALNY/ESC primary attribute the fail to ASO-specific toxicity vs class / tafamidis background?

Round 2 (drill-down):

  1. Exact stabilizer/tafamidis fractions, NT-proBNP and TTR numbers from the simultaneous papers — targeting July “total biomarker null” vs Ionis overall-biomarker claim.
  2. ALNY Q2 primary language on the guide cut and on eplontersen → TRITON-CM — targeting whether −30% is silencer-thesis or orthogonal, and whether ALNY isolated ASO tox.
  3. Timing of the 1,250 → 1,750 expansion vs the 9 July miss — targeting “protocol change because of eplontersen.”

Round 3 (resolve remaining uncertainty):

  1. Post-ESC ALNY / HELIOS-B tafamidis-subgroup primary — targeting whether ESC closed ASO-vs-class.
  2. Confirm no nucresiran outcomes readout and no documented post-ESC protocol rewrite as of 19 Sep 2026.

Anchor source (Grokipedia): skipped — user method “No encyclopedia-as-source.” Noted, not cited.

X sources: not requested (--include-x off).

URLs fetched (13 successful, 4 failed):

Round 1:

Round 2:

Round 3:

Tools used: WebSearch, WebFetch. Grokipedia skipped. x-fetch not used. Generated: 2026-09-19 21:09 UTC

Referenced by