There are bad weeks in pharma, and then there are weeks that force an entire industry to question a foundational assumption. Novartis just had the second kind. Over roughly ten days in early September, the Swiss drugmaker disclosed three separate clinical failures — a CAR-T safety hold, a cardiovascular outcome miss, and a neuromuscular trial halt — erasing more than $40 billion in shareholder value. The market reaction was swift and brutal, but the real story is bigger than one company’s stagger. According to Lauren Marts, speaking on BioCentury This Week, the pelacarsen result in particular raises a question that has haunted cardiovascular drug development for years: when a drug does exactly what it was designed to do biologically, and patients still die at the same rate, what exactly was wrong?
Marts frames the uncertainty in three clean possibilities: “I think at this time we don’t have enough information to determine whether it’s the target that’s the problem, the molecule or the trial design in patient population.” Each carries vastly different implications for the Lp(a) field, for the companies betting billions on it, and for the broader RNA therapeutics space.
The Unraveling of a Perfect Narrative
At the end of August 2026, Novartis was, in Stephen Anson’s words, “one of the best performing and one of the brightest kind of futures pharmas.” The company had risen roughly 20% through the year on optimism about its development portfolio. It had a compelling story: a diversified pipeline spanning cardiovascular, immunology, neurology, and a newly acquired neuromuscular franchise from the $12 billion Avidity Biosciences takeover.
Then came the sequence that broke the narrative.
Approximate DateProgramModalityIndicationOutcomeLate August 2026Rapcabtagene autoleucel (rap-cel)CAR-T cell therapyAutoimmune conditionsThree patient deaths; eight trials pausedEarly September 2026Pelacarsen (Lp(a)HORIZON)Antisense oligonucleotideLp(a)-driven cardiovascular diseaseFailed to reduce cardiovascular events~September 7–8, 2026Del-desiran (HARBOR)Antibody-oligonucleotide conjugateMyotonic dystrophy type 1Failed Phase 3 primary endpoint
The market, notably, shrugged off the CAR-T deaths — shares actually rose 4% on the day that pause was disclosed, buoyed by positive multiple sclerosis data from remibrutinib in the same window. It was the two efficacy failures that broke investor confidence. And among those, the pelacarsen result is the systemic event.
The Lp(a)HORIZON trial was “probably the biggest study that everyone was looking forward to this year,” Anson noted. The genetic evidence linking elevated lipoprotein(a) to cardiovascular outcomes had made Lp(a) one of the most compelling validated targets in cardiology — roughly 90% of Lp(a) levels are genetically determined, untouched by diet and exercise, and about one in five people worldwide carries elevated levels. Analysts had modeled peak annual sales of $3 billion to $6 billion for pelacarsen. The drug did lower Lp(a) substantially, as designed. It simply did not translate into fewer heart attacks, strokes, or cardiovascular deaths.

Three Explanations, Three Very Different Futures
Marts’s framework — target, molecule, or trial design — is not merely a diagnosis. Each possibility reshapes the investment case for an entire field.
The most benign explanation for the rest of the industry is that the problem was the trial itself. Simone Fishburn, the podcast’s editor-in-chief, pushes this angle, drawing an analogy to the amyloid hypothesis in Alzheimer’s disease: the intervention may simply be coming too late. Pelacarsen enrolled patients who had already suffered one cardiovascular event — a secondary prevention population with decades of accumulated vascular damage from lifelong elevated Lp(a). By the time those patients enter a trial, their arteries may already be irreversibly compromised. In that framing, the target is still valid; the timing is the flaw.
The most damaging explanation is that the target itself is wrong. As Marts cautions: “We’ve seen in cardiovascular disease across other indications before that what looks like a good therapeutic target based on pre-clinical and genetic type studies is associated with risk or is associated with the disease but maybe is not actually an actionable target.” Genetic association, in other words, does not guarantee druggability. If that is the case, the entire Lp(a) field — Amgen’s olpasiran, Eli Lilly’s lepodisiran and muvalaplin, and dozens of earlier-stage programs — is built on a premise that may not hold.
The third explanation, specific to the molecule, points to a modality question that the podcast dissects in detail. Pelacarsen is an antisense oligonucleotide, and it achieved up to approximately 80% knockdown of Lp(a) in Phase 2 — a substantial reduction, but meaningfully less than the near-100% knockdown achieved by siRNA competitors. Marts is blunt about what this pattern tells her: “We keep seeing the same story play out when you’ve applied these same two RNA modalities to the same indications. And it’s just you’re often seeing a higher level of knockdown with siRNA than you are with antisense.”
AttributeAntisense (Novartis/Ionis)siRNA (Amgen, Lilly, others)Lp(a) knockdown in Phase 2Up to ~80%Upper 90s, approaching 100%Dosing frequencyMonthlyEvery 6 monthsPhase 3 statusFailed (Lp(a)HORIZON)Ongoing (OCEAN(a), ACCLAIM-Lp(a))
The market has already voted with its feet. Amgen shares fell approximately 9% — roughly $21 billion in market value — on the pelacarsen news, while Eli Lilly ticked down about 2%, wiping out nearly $25 billion. Those are not trivial reactions for drugs that did not fail. They are the market pricing in the possibility that Lp(a) is simply not an actionable target — or that the siRNA advantage in knockdown matters critically.
Fishburn states the stakes precisely: “If the siRNA product achieves the desired outcomes against LP(a) while the Novartis antisense oligo does not, it would be a significant blow to the antisense field versus siRNA.” Amgen’s OCEAN(a) trial, with 7,297 patients and a primary completion date around March 2028, is now the pivotal event for the entire field. If it succeeds, the story becomes about modality superiority and trial design. If it fails, Lp(a) drug development may be over for a generation.
The timing problem compounds the challenge. As Anson notes, you cannot get regulatory approval for primary prevention without outcomes data, and you cannot get outcomes data without running the very secondary-prevention trials that just failed. Statins succeeded because LDL-cholesterol reduction was validated against outcomes before widespread primary-prevention use. Lp(a) now lacks that validation, and the path to getting it is unclear.

The Quiet Hedge
One detail from the podcast that deserves more attention than it will get: Novartis had already hedged its bet. In 2024 and 2025, the company signed two separate deals with Argo Biopharma, a Chinese private company, paying $160 million and $185 million upfront respectively, for siRNA programs targeting cardiovascular disease. One of those deals yielded an Lp(a)-targeting siRNA now in Phase 2, with data expected next year.
Whether that was franchise-building or a quiet hedge against pelacarsen’s failure is impossible to know from the outside. But the outcome is clear: Novartis now finds itself behind Amgen and Eli Lilly in the siRNA race, but still in the game. If the modality question resolves in favor of siRNA, Novartis has a seat at that table. If it does not, no one does.
The Risk Appetite Paradox
The most jarring element of the podcast’s analysis is the contrast between the Novartis failure and what the broader industry is doing simultaneously. Marts’s fourth annual analysis of big-pharma dealmaking — covering 221 deals across 21 top pharma companies from mid-2025 to mid-2026 — reveals a sector that has not retreated into caution but has doubled down on risk.
Deal activity is up versus the prior period, nearly matching levels from two years ago. But the composition has shifted decisively toward what Marts calls “major disruptive potential.” Bispecifics lead all modalities, with T-cell engagers increasingly deployed for autoimmune disease, not just oncology — the logic being to “reset the immune system in the same way that you might with a CAR-T but in a more accessible way.” In-vivo CAR-T continues to rise. Targeted protein degraders and molecular glues round out the portfolio.
In the RNA space, the data is stark: siRNA deals were sharply up, while antisense deals were zero across the entire 221-deal set. That is not a subtle signal. The industry has already made up its mind about which modality it is betting on.
This is a meaningful departure from the bear-market posture of two years ago, when pharmas concentrated on “big opportunities with validated targets.” The current deals carry “relatively little clinical validation” — a deliberate bet that new biology will outperform the old. And that is the paradox: the industry is simultaneously watching a genetically validated, mechanistically sound drug fail in exactly the way its biology predicted it should succeed, and responding by betting on less validated, more disruptive approaches.
The China Shift Nobody Talks About
The dealmaking data also complicates the popular narrative of US decline. Marts’s numbers show Chinese companies now account for roughly the same share of pharma licensing deals as European partners — a significant rise. Critically, it is not a zero-sum shift: “It’s not actually at the expense of the US because the number of deals from US biotech also went up significantly so it’s more like the whole pie grew and more of that was from China rather than farmers are going to China rather than in the US.”
The type of Chinese innovation being licensed has also matured. Two years ago, China deals were “so heavily focused on cancer” and next-generation antibody technologies. Now the modality and disease-area distribution is nearly even — peptides for GLP-1/obesity, RNA therapies, and other emerging platforms.
The striking anomaly is M&A. Despite the licensing surge, there were zero acquisitions of Chinese biotechs during the period, in a year when M&A was otherwise a dominant theme. Jeff Cranmer, the podcast host, recalls asking a biotech executive why Chinese M&A is so rare: “It’s just too much of a headache… just too many hoops to jump through.” The only notable exception remains AstraZeneca’s acquisition of Gracell two years prior.
This gap is now a live policy question. Steve Uden reports that Treasury is considering adding biotech to the CFIUS-style outbound investment restrictions targeting China. Venture capitalists are pushing back, arguing that US and European companies acquiring high-value Chinese assets at relatively low cost is a net benefit to American biotech. The licensing-versus-M&A asymmetry will likely be central evidence in that debate, which is far from settled.
The MFN Carve-Out Game
The episode’s final substantive thread concerns the Trump administration’s most-favored-nation drug pricing deals, where the story is no longer who signed but who did not, and what exactly the carve-outs hide. The White House’s September announcement was notable for its absences — no major US biotech or pharma companies attended.
Uden’s reporting reveals the pressure dynamic. Insmed signed an MFN deal but was not at the White House; its SEC filing discloses it agreed to give state Medicaid programs access to brensocatib at international reference prices. Its absence from the White House event, according to Uden’s sources, reflects the fact that Insmed made no commitment to tie future launches to MFN pricing — a meaningful concession other signatories apparently made.
BridgeBio, which did attend, has stated it expects no future pricing mandates. Its SEC filing reveals a “forward-looking exclusion” for medicines approved exclusively for orphan indications — which conveniently covers its entire pipeline.
The administration has two regulatory hammers: the GLOBE model for Medicare Part B and the GUARD model for Medicare Part D. Companies that signed MFN deals have been promised exclusions from these models, but the scope of those exclusions — whether they cover specific drugs, entire portfolios, or future launches — remains undisclosed.
The Section 232 tariffs, up to 100% on certain drugs and scheduled to take effect September 29, 2026, are the lever that brought foreign companies to the table. Cranmer predicts more tariff-exposed companies will announce MFN deals before that deadline.
The transparency question may resolve itself through politics. Stephen Anson offers a prediction with high confidence: if either the House or Senate flips in the November 2026 midterms, Congress will subpoena and release the MFN deals, revealing details currently unknown about exclusions and portfolio coverage. In his view, the opacity is not likely to survive a change in congressional control.
The Man Who Built the Rules
The episode closes with a remembrance of Peter Barton Hutt, the FDA’s chief counsel from 1971 to 1975 who died at 91. His influence on the regulatory architecture pharma operates within is hard to overstate, and Anson’s tribute frames it precisely: “A lot of the features of FDA law that are taken for granted today really can be traced back to Peter and to things that Peter did.”
The concrete achievements are extraordinary. The adequate-and-well-controlled investigation standard — the foundation of modern drug approval — survived because Hutt won four Supreme Court cases upholding it in a single year, 1973. Four Supreme Court victories in one year is a record few attorneys will ever approach. The off-label doctrine, permitting physicians to lawfully prescribe approved drugs for unapproved uses, traces directly to Hutt, and the motivation was personal: his brother had a serious condition treatable only by an off-label drug, and Hutt assured the physician it was legal, then fought to enshrine that assurance in law. Nutrition labeling on foods was another Hutt championing.
For an industry now navigating the intersection of clinical failures, pricing pressures, and geopolitical friction, the Hutt remembrance is a reminder that the rules shaping pharma’s operating environment were built by individuals making consequential decisions under uncertainty — and that those decisions compound over time in ways their architects could not fully foresee.
The week’s events converge on a single observation about risk appetite. Novartis’s pelacarsen failure is a reminder that even the most rational bets — genetically validated target, mechanistically sound drug, well-powered trial — can fail. Yet the dealmaking data shows the industry is not retreating to safety but leaning further into unvalidated, disruptive biology. That is either a profound insight or a profound mistake, and the distinction will not be clear until Amgen’s readout arrives. That trial, and the congressional subpoenas that may follow the midterms, are the next defining moments. For investors watching the Lp(a) field, the question is no longer whether lowering the biomarker works as a biochemical matter. It is whether genetics, timing, and modality can be aligned to save real patients — and the answer is now further away than it seemed a month ago.