The Basket Trial Is the New Black — and the FDA Hasn’t Caught Up

Pull up the final Phase 1 report for Arrowhead Pharmaceuticals’ zodasiran program, published in Nature Medicine in 2026. What you’re looking at isn’t a conventional dose-escalation study. It’s a single protocol that enrolled patients with heterozygous familial hypercholesterolemia, homozygous familial hypercholesterolemia, severe hypertriglyceridemia, familial chylomicronemia syndrome, and mixed hyperlipidemia — five biologically and clinically distinct populations, stratified by endpoint, running in parallel under one master design. That is a basket trial architecture applied to a cardiovascular metabolic target. And it worked.

Zodasiran delivered an LDL-C reduction of up to 90% in heterozygous familial hypercholesterolemia patients and triglyceride reductions exceeding 70% in familial chylomicronemia syndrome — within the same protocol, same RNAi mechanism, same Phase 1 framework. The data aren’t just compelling for a lipid program. They are a case study in what adaptive, population-stratified basket design can accomplish when sponsors stop treating rare-adjacent disorders as sequentially separate development programs.

But behind those numbers lies a quieter crisis. The regulatory infrastructure for reviewing basket trial data — particularly in non-oncology therapeutic areas — has not kept pace with the operational reality sponsors are now building toward. Three signals, converging in the first quarter of 2026, confirm that a structural mismatch is forming between how trials are being designed and how the FDA is prepared to evaluate them.

The Protocol That Outran Its Guidance

Basket trial design originated in oncology. The FDA’s guidance on master protocols — covering basket, umbrella, and platform trials — was finalized in 2022 and is explicitly anchored to oncology use cases. Read its language on endpoint stratification and you will find a framework built around biomarker-defined tumor subgroups, not around metabolic phenotype heterogeneity in lipid disorders. When Arrowhead designed the zodasiran basket trial to enroll five distinct hyperlipidemia populations with population-specific primary endpoints — LDL-C for statin-intolerant and familial hypercholesterolemia patients, triglycerides for chylomicronemia and hypertriglyceridemia patients — they were operating outside the paradigm that guidance was written to address.

That’s not a criticism of Arrowhead. It’s a structural observation about where adaptive design is heading. The zodasiran protocol used ANGPTL3 inhibition as the single mechanistic thread connecting populations whose clinical presentations, lipid phenotypes, and standard-of-care backgrounds are substantially different. One molecule. Five populations. Five endpoint definitions. One IND. The operational efficiency is undeniable: rather than running five separate Phase 1 programs — each requiring its own IRB review cycle, its own site activation timeline, its own safety monitoring committee — a single basket structure compresses that burden dramatically.

The problem is that FDA’s current guidance on master protocols offers sponsors limited specific direction on how to handle cross-population safety signal adjudication when adverse event profiles differ meaningfully across baskets. In the zodasiran trial, liver-related findings required monitoring protocols that were not identical across all five populations. That is exactly the kind of complexity the 2022 master protocol guidance gestures at but does not resolve — and it’s where the next generation of basket designs in cardiometabolic medicine will run into regulatory friction.

Two More Signals That Name the Pattern

Zodasiran is not an isolated case. It is the most data-rich example of what I’ll call the Metabolic Basket Migration — the movement of basket trial architecture from its oncology home into chronic disease areas where patient population heterogeneity is phenotypic rather than genomic.

Consider what Regeneron and Sanofi achieved with alirocumab across heterogeneous cardiovascular risk populations in the ODYSSEY program. While not a basket trial in the master protocol sense, the ODYSSEY OUTCOMES trial enrolled over 18,000 patients across a range of acute coronary syndrome phenotypes, using a single LDL-C-lowering mechanism to demonstrate benefit across strata. The lesson the field took from ODYSSEY wasn’t the 15% relative reduction in major cardiovascular events — it was that stratified enrollment under a single protocol can generate regulatory-grade evidence across populations that historically would have required separate programs. That logic has now migrated fully into Phase 1 basket architecture.

The second signal comes from the rare disease space. FDA’s guidance on drug development for rare diseases, updated in 2023, explicitly acknowledges the challenge of small, heterogeneous patient populations and encourages sponsors to consider shared natural history data and platform approaches. What it does not do is address how a basket trial with fewer than 20 patients per arm in the rarest populations — as was the case in the familial chylomicronemia syndrome cohort of the zodasiran trial — should be evaluated for dose selection adequacy when each basket has its own PK/PD relationship. That gap is operational, not philosophical, and it surfaces the moment a Phase 2 team tries to use Phase 1 basket data to justify a dose for a registration trial in a single population.

A third signal, less visible but equally important: the FDA’s evolving posture on RNAi therapeutics. Alnylam’s inclisiran program — a PCSK9-targeting RNAi agent approved for hypercholesterolemia — took a conventional single-population development path through Phase 1, 2, and 3. Arrowhead’s zodasiran targets ANGPTL3, a different lipid pathway node, and chose the basket architecture instead. Two RNAi programs in the same therapeutic category, five years apart, using fundamentally different trial designs. That divergence is a signal about where sponsor strategy is moving — and it creates a de facto regulatory precedent gap that the FDA will eventually need to address explicitly, either through updated guidance or through the advisory committee process.

Who Breaks First When the Model Scales

The counterintuitive read on the Metabolic Basket Migration is that technology vendors are better positioned to absorb it than sponsors are. The conventional assumption is that basket trials create complexity that benefits large CROs and integrated sponsors with deep regulatory affairs functions. Look more carefully and the operational stress actually falls on site-level data capture and EDC architecture.

When five patient populations enroll under one protocol but generate five distinct primary endpoint datasets — some LDL-C-based, some triglyceride-based, each with population-specific monitoring schedules — your EDC needs to handle conditional branching logic that most legacy systems were not built to manage cleanly. CDISC SDTM domain mapping for a basket trial with divergent primary endpoints across arms is not a solved problem in most eClinical implementations. Sponsors who assume their existing EDC vendor can configure a multi-basket cardiometabolic trial the same way they’d configure a single-arm rare disease study are going to discover that gap in the middle of a database lock, not before it.

For CROs, the challenge is monitoring. Risk-based monitoring frameworks — built on FDA’s 2023 guidance on risk-based quality management — were designed with homogeneous protocol populations in mind. A basket trial with heterogeneous safety profiles across populations requires that your RBM platform can segment its centralized monitoring triggers by basket, not just by site. Most platforms cannot do this natively. That is a configuration gap that adds months to study startup and, if unaddressed, creates data quality vulnerabilities that will surface in an inspection.

For sponsors building cardiometabolic pipelines around RNAi or antisense mechanisms — Arrowhead, Alnylam, Ionis, Silence Therapeutics — the zodasiran Phase 1 basket result is both validation and warning. The validation: a single protocol can generate enough signal across rare-adjacent populations to justify parallel Phase 2 programs in two or three of the five baskets simultaneously. Ionis, for example, has used stratified development approaches across its lipid and cardiometabolic portfolio to compress timelines. The warning: FDA’s review division for cardiometabolic drugs has not publicly articulated how it will evaluate cross-basket safety data from a Phase 1 program when populations differ in baseline hepatic function, concomitant medication burden, and lipid-lowering treatment history — all of which apply directly to the zodasiran population strata.

FDA’s 2022 master protocol guidance states that “sponsors should consult with the relevant FDA review division early in development to discuss the basket trial design.” That sentence is doing an enormous amount of work. It is, functionally, an acknowledgment that the guidance itself is incomplete — and that the real regulatory framework for any given basket trial will be negotiated in Type B meetings, not derived from published doctrine.

Sponsors who treat that sentence as reassurance are misreading it. Sponsors who treat it as a starting gun for early engagement — submitting detailed basket design rationales in pre-IND meetings before a single patient is enrolled — will build the precedent that eventually becomes the guidance update the field actually needs.

Within 18 months, the first NDA or BLA submission built substantially on a non-oncology basket trial Phase 1 foundation will force FDA’s hand. Either the agency will accept the cross-basket safety synthesis as sufficient for Phase 2 dose selection, establishing a workable precedent, or it will issue a Complete Response Letter that resets the field’s expectations about what basket architecture can carry into registration. The sponsors who have already had those Type B conversations will be protected. The ones relying on the 2022 guidance to do the regulatory heavy lifting will not be ready for what comes back.

References

1. Nature Medicine — “Zodasiran for cholesterol and triglyceride lowering in patients with hyperlipidemia: final report of phase 1 basket trial” (2026)
2. FDA Guidance — Master Protocols: Efficient Clinical Trial Design Strategies to Expedite Development of Oncology Drugs and Biologics (2022)
3. FDA Guidance — Rare Diseases: Natural History Studies for Drug Development (2023)
4. FDA Guidance — A Risk-Based Approach to Monitoring of Clinical Investigations: Questions and Answers (2023)
5. ODYSSEY OUTCOMES: Alirocumab and Cardiovascular Outcomes after Acute Coronary Syndrome (Schwartz et al., 2018)
6. FDA Drug Approval — Inclisiran (Leqvio), Alnylam/Novartis, hypercholesterolemia indication
7. CDISC — Study Data Tabulation Model (SDTM) Implementation Guide, current version