What Semaglutide’s Fentanyl Data Should Mean for How We Design the Next OUD Trial

Buried inside the June 25, 2026 issue of the New England Journal of Medicine is a result that the clinical trial design community should be studying as carefully as any infectious disease team: a pragmatic trial of a 6-month treatment strategy for rifampicin-resistant tuberculosis that generated actionable comparative evidence without the infrastructure burden that typically makes high-quality data collection in resource-limited settings impossible.

The signal here reaches well beyond tuberculosis.

The trial network behind this result, which includes the endTB consortium Médecins Sans Frontières, Partners in Health, and Interactive Research and Development , has spent years building pragmatic, embedded trial infrastructure in settings where standard randomized controlled trial machinery collapses under its own weight.

What they produced is a design object worth examining: a multi-arm, open-label study enrolling patients with RR-TB across multiple countries, generating comparative efficacy and safety data under conditions of genuine clinical heterogeneity, not the controlled homogeneity that makes Phase 3 results routinely irrelevant to the populations most affected by the disease.

The question the rest of the trial design community needs to sit with is whether this model scales beyond infectious disease, and what regulatory systems need to accept before it can.

The stakes are not abstract.

The WHO Global Tuberculosis Report 2025 estimates 390,000 new MDR/RR-TB cases annually and 150,000 deaths attributable to drug-resistant TB in 2024 alone.

A disease at that scale cannot wait for the 18-month site qualification timelines and 40-page eligibility criteria that characterize conventional Phase 3 development.

The Pragmatic Design Architecture

What separates this trial from the standard RR-TB evidence base is not just the 6-month duration target.

The foundational decision was to embed randomization into existing treatment infrastructure rather than erect a parallel clinical research apparatus.

Patients were enrolled and managed at facilities already treating RR-TB under national program conditions, with endpoints defined around outcomes that matter to those programs: treatment success at 72 weeks, culture conversion, and adverse event rates that reflect real prescribing patterns rather than protocol-mandated dose modifications.

This approach directly responds to the methodological gap the FDA identified in its draft guidance on Integrating Randomized Controlled Trials into Routine Clinical Practice, which frames the core problem plainly: highly controlled trial environments generate data that regulators can interpret but clinicians cannot use, while real-world practice generates usable data that regulators cannot trust.

The endTB pragmatic design attempted to satisfy both constraints simultaneously, and the NEJM publication represents the evidentiary output of that attempt.

The comparison point matters here.

The TB-PRACTECAL trial, whose interim results showed 89% treatment success with the BPaLM regimen bedaquiline, pretomanid, linezolid, moxifloxacin , was conducted under more controlled conditions.

Its results were compelling enough to drive WHO’s April 2025 landmark update recommending the all-oral, 6-month BDLLfxC regimen for MDR/RR-TB.

But translating that recommendation into operational reality across high-burden, low-resource settings requires pragmatic evidence, not just efficacy signals from controlled environments.

The NEJM publication fills precisely that gap.

That gap has regulatory consequences no one has fully resolved.

What Regulators Have Not Said Out Loud

The FDA’s expedited approval of pretomanid in August 2019, part of the BPaL regimen for XDR-TB and treatment-intolerant MDR-TB, was built on limited cohort data from the Nix-TB trial conducted by TB Alliance.

The agency accepted that data because the unmet need was catastrophic and the patient population had no alternatives.

What the agency has not done is articulate, with any operational precision, how pragmatic trial data from multi-country, program-embedded studies maps onto its regulatory evidence standards for new indications or label modifications.

The WHO took the sharper position.

Its April 2025 consolidated TB treatment guidelines restructured the recommendation architecture for MDR/RR-TB based substantially on pragmatic and adaptive evidence sources.

That is a normative shift in how an authoritative global health body values evidence generated outside conventional Phase 3 conditions.

The FDA has not made a parallel shift, and that asymmetry creates a real operational problem for sponsors and program implementers trying to use pragmatic trial outputs to support regulatory submissions in the United States.

A cost-effectiveness analysis published in 2025 comparing seven short-course DR-TB regimens in India found that all seven short-course options 6 to 9 months were more cost-effective than the 9- to 11-month standard of care from a

Peptides referenced: Semaglutide.