Cholesterol drug setback casts doubt over multibillion-dollar race

Novartis' late-stage pelacarsen trial lowered Lp(a) cholesterol but failed to significantly improve cardiovascular outcomes, undermining a potential $4 billion-$5 billion peak-sales opportunity and raising doubts over the broader Lp(a) drug class. Novartis shares fell 3%, while Amgen dropped about 5%, Ionis declined 10%, and NewAmsterdam Pharma fell 12% after the update. Analysts said the result increases the burden on Amgen's olpasiran and Lilly's lepodisiran to show that deeper Lp(a) lowering can deliver roughly a 15% reduction in major adverse cardiovascular events.

Analysis

The key repricing is not lost peak sales alone; it is the required-risk-premium reset for cardiovascular outcomes assets built on biomarker validation rather than demonstrated event reduction. AMGN has the largest near-term valuation sensitivity because olpasiran is a more concentrated incremental pipeline driver, while IONS has both economic exposure and a weakened platform proof point. NVS can absorb the asset-specific disappointment operationally, but its pipeline multiple should remain capped until it identifies replacement growth against its post-exclusivity revenue drag.

The decisive variable is whether the miss reflects insufficient target suppression, residual-risk dilution from modern background therapy, or a non-causal Lp(a) signal. If deeper and more durable suppression is necessary, RNAi approaches may retain differentiation, but trials will need either substantially higher baseline Lp(a), cleaner enrichment for recurrent-event patients, or longer follow-up—each increasing development cost and delaying commercial entry by 1-3 years. That makes NAMS particularly vulnerable: its cardiovascular-event narrative may be assigned a lower probability of success even if its mechanism is not directly comparable.

Consensus may be too quick to extrapolate a single failed mechanism to the entire class, but too slow to recognize that a successful future program could address a materially narrower population than prior multibillion-dollar estimates imply. LLY is comparatively insulated because its valuation rests on obesity and diabetes franchises, yet any need to enrich for extreme-risk patients reduces the eventual addressable market and supports lower pipeline-option value. The upcoming detailed dataset is the immediate catalyst: evidence of a favorable trend in patients with very high baseline Lp(a) would reverse the broad class selloff; a flat result across prespecified high-Lp(a) strata would challenge the biological premise materially.

Key Decisions for Investors

• Maintain an underweight AMGN into the full-data presentation; use a 1-3 month bearish put spread rather than outright short after the initial gap. The setup remains negative if high-baseline-Lp(a) and adherent-patient subgroups lack a cardiovascular-event trend; cover if those strata show a credible dose-response relationship or AMGN provides outcome-trial design differentiation that directly addresses the failure.
• Avoid adding to IONS solely on the selloff. Its risk/reward remains asymmetric until the economics of its partnered asset and the mechanism-specific explanation are clear; a further probability-of-success cut would affect both royalty expectations and investor confidence in antisense cardiovascular applications.
• Express relative defensiveness through long LLY / short AMGN over the next 3-6 months, sized modestly. LLY has lower dependence on Lp(a) commercialization, while AMGN faces more direct readthrough; thesis is falsified if the detailed data establish that deeper lowering is likely sufficient and olpasiran shows a clearly superior pharmacodynamic profile.
• Place a risk alert on NAMS rather than initiate after the sharp reaction. Reassess only after management quantifies how its event-driven development path differs from an Lp(a)-specific hypothesis; failure to articulate patient-selection and endpoint differentiation would justify a lower probability-of-success and further multiple compression.