Surrogate Endpoints and Importance to Downstream Decision-Makers

The metabolic dysfunction–associated steatohepatitis (MASH) drug development space is seeing substantial activity, with 2 drugs already approved and several in late-stage development. Both resmetirom and semaglutide were approved under US accelerated approval pathways using histology-based surrogate endpoints. There is a large push by manufacturers, clinicians, patients, and the FDA to move away from liver histology in favor of non-invasive testing (Anania et al., 2025), which will require qualifying non-invasive tests as reasonably likely surrogate endpoints. As teams embark on this work, it's essential to consider not just regulatory stakeholders but downstream decision-makers, HTA agencies, payers, and clinicians. Without accounting for all of their needs, manufacturers risk building an evidence base that only satisfies a fraction of decision-makers.

Why Commercial Teams Need a Seat at the Surrogate Endpoint Table

Surrogate endpoints, "biomarkers or intermediate endpoints intended to substitute and predict for patient-relevant final endpoints" (Ciani et al., Value in Health, 2017), are increasingly the primary endpoints in pivotal trials supporting regulatory approval. That trend has accelerated as sponsors lean on expedited pathways; FDA's accelerated approval program lets drugs for serious conditions reach patients on the basis of a surrogate endpoint "reasonably likely" to predict clinical benefit, with confirmatory evidence generated afterward.

The appeal is obvious. Surrogate endpoints accrue faster than final clinical outcomes, which shortens trials, lowers development costs, and gets promising therapies to patients sooner. In some cases they're also kinder to patients — noninvasive tests like vibration-controlled transient elastography (VCTE/FibroScan), for example, are being developed as alternatives to liver biopsy for assessing fibrosis, sparing patients an invasive procedure.

But that speed comes at a cost: reliance on a surrogate increases the uncertainty of an intervention's true effect on the outcome patients care about.

A surrogate that looks good on paper doesn't guarantee real clinical benefit.

An analysis of FDA cancer drug approvals found that of 36 oncology drugs approved on the basis of a surrogate endpoint, half went on to show no overall survival benefit, and only 14% ultimately demonstrated one (Kim & Prasad, JAMA Intern Med, 2015). That's precisely why validating the relationship between the surrogate and the final endpoint is essential.

The surrogate evidence bar rises from RLSE to validation

FDA recognizes two levels of surrogate endpoint: reasonably likely and validated. Its framework (Anania et al., Hepatology, 2026) lays out three questions that must be answered to qualify a biomarker as a reasonably likely surrogate endpoint (RLSE), the level sufficient for accelerated approval:

• Biological plausibility — is there a clear scientific rationale for why the biomarker sits on the causal pathway of the disease, rather than merely correlating with it?

• Prognostic evidence — does the data show the biomarker predicts clinical outcomes? This is a natural-history question: do patients with worse biomarker values go on to experience worse outcomes?

• Longitudinal disease monitoring / magnitude of change — what magnitude and duration of change in the biomarker predicts a clinically meaningful change in outcomes?

The highest level, a validated surrogate endpoint (VSE), asks a harder question: does a drug's effect on the surrogate predict its effect on the clinical outcome? Validating a surrogate requires randomized treatment data, ideally a meta-analysis across multiple trials, showing that the treatment effect on the biomarker correlates with the treatment effect on the final outcome. The EU's official Joint Clinical Assessment guidance on outcomes states that a correlation of at least 0.85 between a surrogate and the final patient-centred outcome is "described as high" and can be used as a criterion for validation (European Commission, Guidance on Outcomes for Joint Clinical Assessments).

Clearing the RLSE bar and establishing a validated surrogate are not the same accomplishment.

Regulators and HTA bodies are interested in different levels of surrogate endpoint evidence

Regulators and HTA agencies/payers have different remits, and this leads to different needs around surrogate endpoints. Regulators answer: is there enough evidence to conclude a drug

is safe and works, or is reasonably likely to work, for a serious condition with unmet need? HTA agencies and payers ask a different question: is this drug worth reimbursing, relative to what's already available, given its long-term clinical benefit and, in most cases, cost-effectiveness? That question is far less forgiving of RLSE-level evidence, a cost-effectiveness model built on a surrogate endpoint inherits its uncertainty about long-term effectiveness, which compounds over a lifetime horizon. HTA bodies are, in effect, asking for VSE-level evidence.

What it looks like when HTA bodies push back on an unvalidated surrogate

Resmetirom's (Rezdiffra) approval in MASH illustrates this. When FDA granted resmetirom accelerated approval in 2024, it accepted a histology-confirmed RLSE — improvement in liver fibrosis or resolution of MASH. FDA has been explicit, though, that no VSE has ever been established for MASH, and confirmation of clinical benefit remains pending.

NICE's draft guidance on resmetirom (NICE TA11414) pushes back on this RLSE on two fronts. First, it questions the endpoint's relevance to practice: liver biopsy determined treatment eligibility and outcomes in the pivotal trial, but NHS practice would use non-invasive tests like VCTE/FibroScan instead — NICE is effectively calling for a different, non-invasive surrogate to be validated. Second, and more fundamentally, NICE questions the surrogate's validity itself: fibrosis stage is associated with adverse liver-related outcomes, but the evidence shows association rather than causation, and no pharmacotherapy trial has shown that treatment-induced fibrosis regression reduces liver-related events or mortality — RLSE-level prognostic association, but not VSE-level treatment-response validation. This uncertainty carries into the cost-effectiveness model: because the pivotal trial ran only 52 weeks, it had to extrapolate resmetirom's short-term fibrosis effect into lifetime transition probabilities between fibrosis, cirrhosis, decompensation, and death — assumptions resting entirely on the surrogate's association with those outcomes. NICE's committee concluded these compounding uncertainties made it "not possible to determine the most likely cost-effectiveness estimates for resmetirom." (NICE's guidance on resmetirom was in consultation at the time of writing and is subject to change in the final appraisal.)

Other examples show the same pattern. Aclidinium bromide was approved by FDA for COPD based on improved forced expiratory volume (FEV-1), but reimbursement has since been sharply limited, with some payers declining coverage outright and others imposing dispensing restrictions. Belimumab was approved for lupus on a surrogate endpoint but denied reimbursement by several EU national payers (Bognar et al., J Health Econ, 2017).

This limbo — approved but not reimbursed, or reimbursed only on unfavorable terms — is common: roughly 40–50% of assessments by Germany's G-BA (Wehage, et al. 2024) and France's HAS (Lanecka, et al 2024) conclude a drug offers no added clinical benefit over existing options, and most of NICE's positive oncology recommendations are conditional on price concessions or restricted to narrower populations (Dickson, et al 2018).

Considerations for sponsors: RLSE is a checkpoint, not a finish line, and clinical development can't own it alone

Sponsors should account for the needs of different stakeholders when developing surrogate endpoint strategies — this can't be executed by clinical development in isolation. Clinical development alone can clear FDA's bar, but it can't judge what HTA agencies will accept downstream; that's exactly why HEOR and Market Access need to be in the room from the start, shaping decisions on trial outcomes, validation evidence packages, and scientific advice questions.

1. Don't mistake the RLSE benchmark for the finish line. Clearing the RLSE benchmark and winning accelerated approval is a real accomplishment, but it may not be sufficient for optimal reimbursement, and treating it as "done" is a strategic error. Commercial and market access teams that aren't closely tracking the surrogate validation work can inherit an evidence base that falls short of what HTA bodies need.

2. Treat the RLSE period as a cross-functional evidence-generation window, not a pause. The gap between RLSE and VSE is a defined body of work, the confirmatory outcomes trial, treatment-effect correlation data, real-world evidence, and generalizability data HTA bodies are asking for, and cross-functional teams should plan for it by designing confirmatory trials and RWE programs that double as surrogate validation studies.

3. Bring HTA agencies — and the internal teams who own that relationship — into the conversation early. Just as sponsors engage FDA directly through Type C meetings to move a biomarker from RLSE to VSE-level evidence, the same proactive engagement belongs on the HTA side: HEOR/Market Access teams should pursue scientific advice meetings, early dialogue on validation strategy, and transparency about the confirmatory evidence timeline.

None of this is quick. Validating a surrogate to the standard HTA bodies actually want — treatment-effect evidence, not just prognostic association — takes years, often the same years required to complete the confirmatory trial FDA already mandates. RLSE-based accelerated approval buys time and market access — it doesn't buy reimbursement, because the evidence that satisfies a regulator isn't automatically the evidence that satisfies a payer.

This is exactly the kind of cross-functional planning gap we help sponsors close. At Pedestal Health, we work with clinical, commercial, and market access teams early to design real-world evidence programs that can serve as surrogate validation studies, generating the data that regulators and HTA bodies need before the confirmatory trial reads out. If you're navigating this gap on a current program, we'd welcome the conversation.