Five.
Five regulatory bodies — the European Medicines Agency, the UK's MHRA, the European Pharmacopoeia, the transatlantic TATFAR taskforce, and Germany's BfArM — published phage therapy regulatory frameworks between March 2024 and April 2026. More regulatory convergence for bacteriophage medicine than in the previous hundred years combined.
Tomorrow, the EMA's public consultation on its draft guideline for phage therapy medicinal products closes. The six-month comment period on EMA/CHMP/BWP/1/2024 ends April 30, 2026. A final guideline is expected by Q3 2026.
The question that hangs over all five frameworks is the same: who is left to use them?
The Convergence
| Body | Date | Scope | Key Feature |
|---|---|---|---|
| EU Pharmacopoeia | Mar 2024 | Quality standards | Chapter 5.31 — first harmonized quality criteria for phage APIs. In force Jan 2025. |
| Germany (BfArM) | Sep 2024 | National law | Medical Research Act amends AMG. Creates two pathways: industrial + magistral pharmacy (no marketing authorization required). |
| MHRA (UK) | Jun 2025 | Regulatory guidance | Covers licensed + unlicensed use, GMP, personalized treatments. Classifies phages as Biological or ATMP. |
| TATFAR | Dec 2025 | Transatlantic perspectives | Nature Comms paper integrating EC HERA, EMA, FDA, MHRA, Health Canada, NIH views. Identifies regulatory gaps and manufacturing bottlenecks. |
| EMA | Oct 2025 → Apr 2026 | Quality guideline | Draft guideline on quality aspects. Two-tiered seed lot system, full genome NGS, lytic phages only, mediated transduction screening. Consultation closes Apr 30. |
The table looks like progress. It is progress. But zoom out and a darker pattern emerges: these frameworks arrived just as the companies they were designed for collapsed.
The Graveyard
In June 2025, PHAXIAM Therapeutics was judicially liquidated by the Lyon Commercial Court — €150 million in debt, €3.6 million in cash, no acquirer. The company had treated 140 patients. It was formed from the merger of Pherecydes Pharma (founded 2006, one of the oldest phage companies in Europe) and ERYTECH's clinical assets. Two decades of phage development, delisted.
In February 2026, Seres Therapeutics suspended investment in SER-155, its microbiome therapy for AMR decolonization in transplant patients — Breakthrough Therapy designation and all. Thirty percent of the workforce was cut. Cash through Q3 2026 only.
BiomX discontinued its most advanced phage program (BX004 for CF-related infections) in December 2025 after a Data Monitoring Committee flagged adverse events. Its Israeli subsidiary is insolvent. Cash as of September 2025: $8.1 million. A new CEO is attempting a strategic reset around its defense-funded diabetic foot program (BX011, $40 million DHA/Navy).
Venatorx — not a phage company but a cautionary parallel — had its equipment auctioned in October 2025. Four hundred lots. Its only surviving asset: ceftibuten-ledaborbactam, licensed to Basilea for up to $325 million in milestones.
The pattern is not subtle. The pharmaceutical pathway for phage therapy has produced zero FDA-approved products in over thirty years of modern development, while consuming hundreds of millions of dollars and killing nearly every company that attempted it.
The Survivors
Every surviving phage clinical program of significance shares one feature: government funding.
S. aureus bacteremia phage cocktail. QIDP + Fast Track. Phase 2a: 88% vs 58% response (p=0.047), 2.7 vs 9.3 days to resolution. Phase 3 initiation H2 2026. ~450 patients, superiority design.
Funding: $26.2M DoD. Cash: $14.1M against $150.3M liabilities. Going concern.
CRISPR-Cas3 engineered phage for UTI. ELIMINATE trial enrolling 288 patients.
Funding: $93M BARDA (up to $152M total).
CRISPR-Cas-armed phage for gut E. coli decolonization. Phase 1 results in Lancet Microbe (Mar 2026) — first randomized, placebo-controlled CRISPR antimicrobial trial in humans. Phase 1b >50% enrolled, 8 US centers, 24 patients.
Funding: CARB-X. FDA FastTrack.
Phage cocktail for diabetic foot infections. Strategic pivot after BX004 failure.
Funding: $40M DHA/Navy. Company cash: $8.1M (Sep 2025).
The list is short. Every entry carries a government contract, a defense grant, or a CARB-X award. The one company closest to a commercial product — Armata, with AP-SA02's Phase 3 initiation planned for the second half of this year — carries going-concern language in its filings. A $26.2 million Department of Defense contract is what keeps it alive.
This is the paradox the five regulatory frameworks must now confront: they have defined how to approve phage products, but the commercial model that would bring those products to market has been tested to destruction.
The Other Road
While companies died and regulators drafted, three European countries quietly built something different: public infrastructure for phage therapy.
Belgium: the clinical model
Since 2018, Belgium has permitted phage therapy through magistral preparations — compounded medicines prescribed by a physician and prepared by a pharmacist, exempt from marketing authorization. The Queen Astrid Military Hospital in Brussels has treated over 100 consecutive patients this way. Clinical improvement: 77.2%. Eradication: 61.0%. Vésale Bioscience, a Belgian spin-off, received a €1.8 million EIC grant for PhageDiag — an automated phagogram that tests 96 phages against a patient's bacterial isolate in two to three hours. The diagnostic bottleneck is being addressed alongside the therapeutic one.
France: the manufacturing model
PHAG-ONE, France's first public GMP phage platform, is building phage banks for S. aureus, S. epidermidis, E. coli, and K. pneumoniae. Its purification process — DNA nuclease treatment, tangential filtration, ultrafiltration — is designed for compliance with Good Preparation Practices. The project was supported by France's National Research Agency. Its ambition: a European reference model for scalable public production. Results will be presented at the Targeting Phage Therapy 2026 congress in Valencia on June 9-10.
Germany: the regulatory template
The Medical Research Act (Medizinforschungsgesetz), adopted by the Bundesrat in September 2024, amended Germany's pharmaceutical law to create two explicit pathways for phage products: standardized industrial manufacturing under full marketing authorization, and magistral pharmacy formulations exempt from it. The PhagoFlow research project at Hannover Medical School has treated patients with personalized phage preparations for Pseudomonas aeruginosa wound infections since 2023, funded by Germany's Joint Federal Committee innovation fund. Thirty-three patients treated at Hannover since 2015, thirty-one successfully — a 94% response rate in critically ill patients.
Belgium built clinical delivery. France built manufacturing. Germany built the regulatory template. Together, they constitute a complete alternative to the pharmaceutical pathway — one that doesn't require any company to survive.
The Moving Target
There is a structural problem that none of the five regulatory frameworks has fully solved, and it sits at the heart of why phage therapy is different from conventional drug development.
Bacteria evolve resistance to phages. When they do, the phage cocktail must change. But a marketed pharmaceutical product — approved through the standard pathway — cannot easily change its composition. The regulatory architecture assumes a fixed product. Phage therapy requires a moving one.
A comprehensive Frontiers review published in 2025 identifies this as the fundamental tension: the industrial pathway (fixed composition, mass production, marketing authorization) clashes with the biological reality of phage-bacteria coevolution. The magistral pathway (personalized, small-batch, physician-directed) accommodates biology but limits scale.
Germany's two-pathway approach is the most honest acknowledgment of this tension. Belgium's eight-year magistral experience is the most battle-tested. The EMA guideline closing for consultation tomorrow stakes out the industrial position: full genome NGS for characterization, a two-tiered seed lot system for consistency, screening for mediated transduction to prevent gene transfer, and potency assays aligned with the EU Pharmacopoeia. It is rigorous. It is appropriate for standardized products. And it may not match how phage therapy actually works in practice — where the pathogen dictates the treatment, not the manufacturer's catalog.
What Is Being Built
The convergence is real. For the first time in the century-long history of phage therapy, regulators across multiple jurisdictions are saying the same thing at the same time: phages are medicines, they need quality standards, and there are pathways to bring them to patients.
But the convergence arrives into a landscape where the pharmaceutical model has been tested and has failed — not because the science is bad (AP-SA02's 88% response rate, Chan's 10,000-fold bacterial reduction in CF patients, Belgium's 77% clinical improvement over 100 patients), but because the economics are identical to the ones that killed antibiotic companies: the drugs that work best are the ones the market values least.
The frameworks now exist. The science is accumulating. Armata may reach Phase 3 this year. SNIPR may complete its CRISPR antimicrobial Phase 1b. Locus may report ELIMINATE data. But if the last five years have demonstrated anything, it is that clinical success in phage therapy does not translate into commercial survival.
The five frameworks were built for a pharmaceutical industry. The three infrastructure models — Belgium, France, Germany — were built for something else: a world where phage therapy is a public service, not a product.
The companies that tried the first road are mostly dead. The countries that are building the second are mostly succeeding. The frameworks close for comment tomorrow. The question they cannot answer is which road we are actually on.