At ARDD 2026, senior FDA officials sketch how a therapy targeting aging might one day reach approval, revealing that longevity is a priority.
The longevity field has spent years asking what the FDA thinks of aging as an indication. On the first morning of ARDD 2026 at Harvard, it got the closest thing yet to an answer – not a pathway, exactly, but a candid map of the terrain from the people who would have to approve whatever comes down the line.
Moderated by Andrew Brack, Program Manager for Proactive Health at ARPA-H, the session – Matching Clinical Trials of Therapeutics & Regulatory Mandates – brought together four senior FDA figures: Jeffrey Siegel, Director of the Office of Drug Evaluation Sciences at CDER; Justin Penzenstadler, Associate Director (Acting) FDA, Lowell Zeta, Deputy Commissioner for Strategic Initiatives (FDA); and Steven Kozlowski, the agency’s Chief Scientist.
The headline came roughly two-thirds of the way in, and drew applause: the FDA is preparing a new focus areas of regulatory science document, and aging and longevity will be one of its topics.
Start narrow, go broad
Brack opened with what he called the biggest framing question in the field: what is the most realistic regulatory path for a therapy that targets aging biology rather than a single disease?
The panel commented: decades of basic research have identified general mechanisms of aging, and blocking some of them has extended lifespan in experimental models by as much as 30% – with newer studies hinting that some aging processes might even be reversed. Translated into people, the implications would be considerable. But there are two distinct routes to proving it. The first builds on the fact that aging drives the prevalence of many chronic diseases, from Alzheimer’s to cardiovascular disease: show that a treatment acting on a general mechanism benefits one of those conditions, then a second and a third (disease), and the case that the mechanism itself is doing the work begins to accumulate.
The second is more ambitious – define a patient population subject to age-related deterioration across domains such as cognitive decline, frailty, and loss of hearing and vision, find an outcome measure that captures that deterioration, and show in a controlled trial that treated patients fare better than untreated ones.
Asked to put some meat on that, Penzenstadler said he expects the earliest approvable trials to look at the prevention of age-related comorbidities or mortality. Alongside those hard endpoints, he anticipates sponsors – and, he suggested, many in the room – collecting intrinsic capacity and other clinical outcome assessments, until enough data accumulates to tie those measures rigorously to survival. He framed it as a typical arc for a therapeutic area, reaching for his own background in diabetes: the first insulin trial was effectively a single, highly enriched patient – a child on their deathbed – with mortality as the endpoint. Geroscience, in his view, will follow a similar path: enriched populations and hard outcomes first, a few wins, and then validated measures that track mortality and allow shorter trials.
A ‘G factor’ for aging
Kozlowski came at the question as a scientist. If aging is a shared risk factor across many chronic diseases, he asked, can that shared component be quantified? He drew an analogy with Spearman’s G factor in psychology – the idea of a single general factor underlying performance across different cognitive tasks: by the same logic, some fraction of many chronic diseases is probably attributable to a shared aging factor, and the better it can be measured, the more naturally it fits into how studies are designed. That, he said, is an important regulatory science goal, and one that better use of longitudinal studies could help to answer.
His ideal is a shared endpoint – pre-competitive, qualified and broadly available – that can serve as a bridge. Its attributes would be regulatory acceptance, feasible trial designs that make development worth doing, a standard against which novel measures can be judged, and a route to accessibility, including the reimbursement that gives developers an incentive to build products in the first place.
Benefit-risk and the repurposing trap
From a policy perspective, the panel questioned how much new guidance the field actually needs to make progress, pointing instead to cross-functional work across the agency on the tools and endpoints that capture how patients feel and function – and to cardiometabolic drugs as a likely testing ground for multimorbidity endpoints.
Primary prevention in healthy people is a different proposition from secondary prevention: treating someone for 30 years to derive a modest benefit demands careful thought. That is why trials in older people, with multimorbidity endpoints, make sense – the treatment horizon is closer to a decade than several, the endpoints are hard, and the benefit-risk calculation is more straightforward.
The panel also flagged a trap for anyone repurposing drugs. SGLT2 inhibitors and GLP-1 receptor agonists are already proven to reduce major adverse cardiovascular events. Run a longevity trial with them and, unsurprisingly, people may live longer – but for reasons already explained by known clinical effects. Sponsors can guard against that in two ways: careful eligibility criteria, and endpoints spanning multiple domains – sensory or functional measures, for instance – that sit outside a drug’s established benefits.
From prognostic marker to surrogate endpoint
With many in the audience generating digital measures and biomarkers, discussion turned to the FDA’s qualification programs – and, in effect, to a primer on the evidence ladder. The key distinction is between a prognostic biomarker, which stratifies people by their risk of developing a disease, and a surrogate endpoint, which is also prognostic but changes with treatment – and that change reflects clinical benefit downstream.
Getting from one to the other is a staged process. Epidemiological evidence must first link the biomarker to later clinical outcomes; longitudinal studies then need to show that change in the biomarker tracks change in risk; and finally, therapeutic trials must show that a treatment-driven change in the biomarker predicts the clinical outcome.
Once proven, a validated surrogate can stand in for clinical endpoints. Short of that, accelerated approval offers a route for serious conditions with unmet need, using a surrogate that is reasonably likely to predict benefit, where the evidence is strong but incomplete. Predictive biomarkers can also be used to enrich trial populations.
The panel was frank that developing surrogates is very hard, particularly for common chronic diseases where clinical evidence is difficult to generate. Mechanism already plays a role in regulatory decisions, but in a rough way – as biological plausibility – and the way companies present it is unstructured, ranging from sophisticated mapping to little more than a slide of boxes, arrows and selected references. A standardized framework for weighing mechanistic evidence would never be the whole story, but it could make that evidence more convincing – or reveal it as less so.
Aging makes the agenda
Delegates were told that the agency is planning to issue a focus areas of regulatory science document – high-level principles on the areas most important to the FDA – and that aging and longevity will be one of its topics.
The announcement drew applause. A timeline was harder to pin down: pressed on dates, the panel noted it had been some time since the last edition. The work was framed within the Commissioner’s office’s broader public health priorities, with an emphasis on giving sponsors predictability.
It is a modest sentence in a long session, but its significance is not. A regulator does not write a research agenda for a field it intends to ignore.
What the regulators want from the field
So, what can the community do to help?
First, form a pre-competitive consortium to work out a registry of biomarkers to include in commercial trials intended for drug approval, with trial designs that allow results to be pooled back into a biomarker validation package, improving continually in a revolving-door fashion. Second, establish recurring meetings between regulators, academia and industry; therapeutic areas that meet regularly with the agency tend to benefit from the structure and from hearing each other out.
On misconceptions, the panel acknowledged that the FDA is sometimes seen as a hurdle to clear – and that it can look like one. But the agency regards aging as an important problem and a huge opportunity, and getting it wrong would hurt the field. The cautionary tale offered was anti-arrhythmic drugs: the assumption that suppressing irregular heartbeats would prevent deaths turned out to be wrong, and set biomarkers back for a long time. The fundamental principles of how the FDA judges evidence are sound and will not change.
Another point was simpler: the FDA is data-driven. Hypothetical questions are very hard to answer, but when sponsors bring data, the agency takes notice. Longitudinal studies that systematically show how aging affects specific outcomes would let developers state exactly what a trial is intended to demonstrate.
The session closed with practical advice. Sponsors asking at a Phase 1 meeting whether a given Phase 3 result will secure approval will hear the agency’s famous answer: it’s a review issue. The best way to get feedback on trials and endpoints is to submit an IND and a protocol through formal channels. For a field trailblazing a new indication, the message was to avoid over-specific questions for now, ask questions grounded in data – and expect the FDA to work through the problem alongside them.
The ball is in the field’s court
Nobody on the panel promised an aging indication; what they offered instead was more useful: a recognizable route – hard outcomes first, then validated measures – a frank account of where the evidence bar sits, and a standing invitation to bring data rather than hypotheticals. The FDA has now said, publicly, that aging belongs on its scientific agenda. Whether the consortia, shared endpoints and longitudinal datasets materialize to meet it is, as the panel made clear, not the regulator’s job.
Read more of our on-the-spot coverage of ARDD 2026 – CLICK HERE.
