An inherited cholesterol-like particle raises heart-attack risk and has no treatment. Several drugs now lower it; the question is whether that prevents heart attacks.
40%likely
Well sourcedNext check
Digging in
Why it is different
The edge is not whether it works. It is that the market will read the coming trial result as a final verdict when it is only a halfway look.
Why it matters
It affects roughly one person in five and there is nothing to offer them today.
Next check · 9 Nov 2026
Is the Lp(a)HORIZON trial presented in a headline session at the American Heart Association meeting in November 2026?
How we scored it ·
from how similar cases went
If right, how big5/5How overlooked3/5Upside vs downside2/5Source trust5/5Independent sources3/5
The exact wording we score
One of three registered hard-MACE outcomes trials — pelacarsen HORIZON (NCT04023552), olpasiran OCEAN(a) (NCT05581303), or lepodisiran ACCLAIM-Lp(a) (NCT06292013) — hits its primary MACE endpoint with statistical significance and that same therapy gains approval for cardiovascular risk reduction by 31 Dec 2030. Resolves YES only if both legs hold for one named therapy; NO otherwise.
PROOF
Evidence and provenance
The links below are the evidence recorded when this forecast was scored. Each note states source quality or bias; a citation is not an endorsement.
Recorded sources · 5 citations / 3 independent origins
2026-07-28 | sell-side: success remains the modal view but timeline risk is recognised; mainstream press: trial delay and management hedging are visible; priced market: no clean single-asset contract identified; specialist: cardiovascular debate now actively questions absolute-event reduction | E3
Revision history4 entries
Probabilities and interpretations are never silently overwritten. This is the append-only record of what changed and why.
(3rd pass) — No score change. Noted in the claim section that this signal was already correctly class-level — a bounded, named three-asset disjunction — and became the template when the other four drug signals were raised to match their edges. Its defect was in the conjunct (approval timing), not the disjunct.
(2nd pass) — P 52→40, edge 2→3, and the Harvest demotion is REVERSED. Two errors from this morning's pass. (1) The probability chain in this file is P(biology) x P(power) x P(approval-by-2030 timing ~90%). The v2 log destroyed that 90% — it found the end-2026 event is a 75%-of-events interim whose modal NO is 'continue to full accrual in 2028', pushing filing to ~2029 — but only moved P from 57 to 52. The term that changed was never recomputed. Re-running this file's own formula with a defensible timing factor (~70%): 0.60 x 0.85 x 0.75 ≈ 0.38-0.40. (2) The auto-Harvest rule fired on a stale edge score. Edge was cut to 2 on 2026-07-23 because the readout looked like a telegraphed consensus binary. Five days later this file discovered that the consensus framing is wrong — the market prices a clean binary that is actually an interim with a live continue-to-2028 branch. That is a non-consensus, disconfirming, structural insight: edge went UP, and the generator demoted the signal to Harvest on the same day. Edge restored to 3, status → investigating. Methodology consequence: E<=2 now FLAGS for harvest review instead of silently demoting — see METHODOLOGY.md v2.1.
v2 rescore. P 57→52 on two findings, and auto-demoted to 🌾 Harvest by the E<=2 rule (edge fell 3→2 on 2026-07-23; v1 left it ranking actively anyway). (1) The ESC 2026 window is closed: the ESC Hot Line programme, published 2026-07-08, contains no Lp(a) trial across all 46 slots (the lipid slot is SHASTA-3/4, plozasiran). This file's 'ESC 2026 as an earlier late-breaker window' is falsified; AHA 2026 is the only remaining 2026 venue. (2) The end-2026 event is an interim, not a final: on the Q2-2026 call management described a readout at 75% of events with stopping criteria, and said that if criteria are not met the trial continues to full accrual in 2028. The modal NO outcome is therefore not 'miss' but 'continue to 2028' — which alone pushes filing to ~2029 and approval to the edge of the horizon. This file's clean-2026-binary framing understates that middle branch. Date correction: AHA 2026 is 6-9 Nov, not 7-9. Registry: NCT04023552 Active-not-recruiting, n=8,323, primary completion 2026-06-30 est. — that date has now passed with no announcement.
created, then deepened same day: HORIZON confirmed reading out end-2026 (near-term binary), management hedge noted, correlated-risk refined; edge→2, probability→57, status→investigating.
Method challenge
Adversarial review record
Admitted after revision
Structured internal role review tied to this frozen claim. It is not independent human peer review. Independent specialist review not yet performed
clinical-domain specialist
biostatistics and trial-design reviewer
regulatory reviewer
forecasting-method reviewer
Strongest specialist challenge
Large biomarker reductions do not guarantee cardiovascular benefit, and event-driven trials can be underpowered if blinded event rates run low. Approval timing is a second hurdle that must be met by the same therapy, not a class-level readout plus another asset's approval.
Outside view
The reference class includes failed lipid biomarker interventions as well as successful LDL pathways. Three programs improve the chance of scientific validation, but their causal biology and regulatory risks are strongly correlated.
Causal rival
Lp(a) can be a risk marker without the amount or timing of pharmacologic lowering producing enough MACE reduction. Background therapy, endpoint composition, population enrichment, and absolute event rates can explain a statistical miss despite target engagement.
Measurement risk
A qualifying YES requires pre-specified primary MACE significance and an FDA cardiovascular-risk-reduction approval for that exact drug by 2030. Interim continuation, subgroup success, biomarker approval, or filing acceptance is insufficient.
Residual risk
ClinicalTrials.gov dates are sponsor-entered and can precede public topline or approval by years. The 40 percent estimate remains particularly sensitive to HORIZON timing and to whether a narrow miss preserves confidence in deeper siRNA approaches.
COLDTELL