A physician in Los Angeles is looking at a stool culture. Shigella sonnei. Resistant to ampicillin, azithromycin, ceftriaxone, ciprofloxacin, and trimethoprim-sulfamethoxazole. Every oral antibiotic she could prescribe — gone. The patient has bloody diarrhea and fever but is not sick enough for hospitalization. Not sick enough for IV meropenem. Just sick enough that they need a pill that works.
There is no such pill. Not one approved by the FDA for this infection.
This scenario is no longer hypothetical. It is happening in American cities right now, and it is the leading edge of a convergence that the infectious disease community has been slow to name: three of the world’s most common bacterial diseases are independently losing their oral treatment options at the same time, through different resistance mechanisms, in different geographies, with the same drug — azithromycin — failing as the common last line in all three.
The convergence
The three diseases — shigellosis, typhoid fever, and gonorrhea — collectively cause more than 200 million infections per year, primarily in children and young adults. They are mucosal and enteric infections. They happen in the community, not the hospital. And their defining clinical characteristic is that oral treatment is not a convenience — it is often the only treatment available, because the patients who carry these infections live in settings where IV therapy, or even hospitalisation, does not exist.
| Drug class | Shigella | S. Typhi | N. gonorrhoeae |
|---|---|---|---|
| Ampicillin | Lost | Lost | Lost |
| TMP-SMX | Lost | Lost | — |
| Fluoroquinolones | Lost | Lost | Lost |
| Third-gen cephalosporins (oral) | Lost (XDR) | Lost (XDR) | Failing (0.8→5%) |
| Azithromycin | Lost (XDR) | Failing | Failing |
| Approved oral options remaining | 0 | 0–1 | 2 (new) |
For gonorrhea, zoliflodacin and gepotidacin were both approved in December 2025. For XDR typhoid, azithromycin remains the only oral option but resistance is spreading. For XDR Shigella, there are zero FDA-approved oral treatments.
The table tells the story that individual disease reports cannot: the same drug classes are failing simultaneously across all three pathogens, and azithromycin — the drug that clinicians reached for when everything else failed — is the final row turning red in each column.
Shigella: zero
The CDC reported in April 2026 what surveillance had been warning about for years: XDR Shigella rose from 0% of US isolates in 2011 to 8.5% in 2023 — 510 of 16,788 isolates resistant to all five oral drug classes. The demographic profile is striking: 86.2% adult males, 46.6% with HIV co-infection. But the resistance is not confined to any population.
In January 2026, researchers at UCLA identified a novel XDR S. sonnei genotype in Los Angeles carrying blaDHA-1, an AmpC-type beta-lactamase that confers ceftriaxone resistance through a mechanism entirely different from the CTX-M genes that drove earlier XDR outbreaks. Whole-genome sequencing showed close genetic relatedness between cases with no identified epidemiological link — silent community transmission of a strain that independently evolved XDR through a novel pathway. Two kinds of XDR Shigella are now circulating simultaneously in one American city, having arrived at the same phenotype through different molecular roads.
The global numbers are worse. Shigella causes an estimated 188 million cases and 164,000 deaths per year, overwhelmingly in children under five in sub-Saharan Africa and South Asia. These are children who will never see a hospital, let alone receive IV meropenem. For them, the oral treatment desert is not a clinical inconvenience. It is a death sentence written in resistance genes.
Typhoid: the last drug standing
Azithromycin is the last effective oral antibiotic for XDR typhoid. It is failing.
A systematic review published in 2026 documented the emergence of azithromycin-resistant Salmonella Typhi across Southeast Asia, driven by acrB mutations and mph(A) acquisition. Pakistan, where XDR typhoid first emerged in 2016 — resistant to ampicillin, chloramphenicol, TMP-SMX, fluoroquinolones, and third-generation cephalosporins — has seen 7.3% of its XDR strains acquire azithromycin resistance. In Bangladesh, the numbers from tested isolates are higher. In India, mph(A)-mediated resistance has emerged independently, suggesting parallel evolution rather than a single spreading clone.
The worst-case scenario is no longer theoretical: in 2024, the first NDM-5 carbapenem-resistant S. Typhi was isolated from a German patient who had travelled to India. A single organism carrying both XDR and carbapenem resistance. If that genotype stabilises and spreads, there will be zero antibiotics — oral or intravenous — to treat it.
Typhoid fever kills approximately 100,000 people per year and sickens 11 million, mostly in South Asia. Typhoid conjugate vaccines are powerful — Gavi has supported the vaccination of over 130 million children — but vaccines prevent new infections. They do not reverse resistance already circulating in bacterial populations. The 7.3% azithromycin-resistant fraction will not shrink because more children are vaccinated. It will persist, waiting for each unvaccinated case, each traveller, each immunocompromised host.
Gonorrhea: two new drugs, already threatened
Gonorrhea is the partial exception. In December 2025, the FDA approved two new antibiotic classes — zoliflodacin and gepotidacin — the first new gonorrhea treatments in decades. GARDP is rolling out access across 160+ countries.
But the resistance clock is already running. A gyrB D429N mutation conferring 16–32-fold increased zoliflodacin MICs, with cross-resistance to gepotidacin in parC D86N backgrounds, has been identified in circulating strains. Ceftriaxone resistance in N. gonorrhoeae climbed from 0.8% to 5% globally between 2022 and 2024. In Vietnam, it reached 32.4%. The two new drugs bought time. How much time depends on deployment strategy — and the history of gonorrhea treatment suggests that single-drug regimens collapse within a decade.
Gonorrhea demonstrates what shigellosis and typhoid are about to learn: even when new drugs arrive, resistance does not reset. It accelerates.
Three diseases. Three independent resistance trajectories. One shared destination: the end of oral antibiotic treatment for community-acquired infections that, together, sicken 200 million people a year.
Three pills that might work — none approved for the job
The cruelest feature of the oral treatment desert is that potential rescues exist. They are sitting in pharmacies, approved for other indications, with published evidence of activity against enteric pathogens. But none is approved, registered, or trialled for the infections where they are desperately needed.
Tebipenem — the oral carbapenem that has never treated an enteric infection in a trial
If the FDA approves tebipenem pivoxil hydrobromide by its June 18, 2026 PDUFA date, it will be the first oral carbapenem available in the United States — for complicated urinary tract infections. Not for shigellosis. Not for typhoid. The PIVOT-PO Phase 3 trial, run by GSK after licensing from Spero Therapeutics, was stopped early for efficacy, demonstrating non-inferiority to IV imipenem-cilastatin.
But the enteric data exists. Published pharmacodynamic modelling shows MICs of ≤0.62 μg/mL against XDR S. Typhi. Paediatric pharmacokinetic studies achieved 92.4% target attainment for shigellosis dosing. A shigellosis trial in Bangladesh (NCT05121974) has been registered. Yet no enteric fever trial exists for a drug that could be the only oral treatment for infections killing hundreds of thousands of children.
Tebipenem was approved in Japan in 2009. It has been used there for 17 years. Paediatric granule formulations exist. The compound is not experimental. It is a regulatory orphan — proven in one geography, filed for one indication in another, and unavailable for the infections where it could have the greatest impact on mortality.
Fosfomycin — zero resistance in 510 XDR Shigella isolates
The CDC’s own data contains a remarkable signal buried in a resistance table: of the 510 XDR Shigella isolates tested, exactly zero showed fosfomycin resistance. In a landscape where every oral drug has failed, fosfomycin remains universally active.
Clinical evidence is thin but real. A California case report documented clearance of XDR S. flexneri 2a with oral fosfomycin (3g every 48 hours, three doses). A UK protocol uses fosfomycin as oral step-down after IV carbapenem for XDR Shigella. Fosfomycin is FDA-approved — for uncomplicated UTI only. It is not approved for shigellosis in any country.
There is a caveat. S. sonnei isolates from poultry in some surveillance studies have shown fosfomycin resistance, raising the spectre of agricultural co-selection — a pattern this blog has documented in other contexts. The window of universal susceptibility may not last.
Pivmecillinam — recently FDA-approved, never studied for shigellosis
During the UK’s XDR S. sonnei outbreak (72 cases), all isolates were susceptible to pivmecillinam. The drug was recently FDA-approved for uncomplicated UTI, making it available in the US for the first time. But no shigellosis trial data exists. No enteric fever trial data exists. The compound is a prodrug of mecillinam, a PBP2-selective antibiotic used in Scandinavia for decades, with a well-characterised safety profile. It sits on pharmacy shelves, active against XDR Shigella, approved for a different infection.
Three oral drugs. All active against XDR enteric pathogens in laboratory testing. None approved for the purpose. None being developed for the purpose with any urgency. The gap between what exists and what is accessible is not a scientific gap. It is a regulatory and commercial gap — the kind that kills people who cannot be hospitalised.
The missing target product profile
In March 2026, the WHO published updated target product profiles for antibacterial agents — the documents that tell the field what to build. The profiles focused on bloodstream infections, pneumonia, and meningitis. There was no TPP for enteric pathogens.
No TPP means no signal to industry. No signal means no trials. No trials means no approvals. The 200 million cases of shigellosis, typhoid, and gonorrhea — mucosal and enteric infections that disproportionately affect children, travellers, and populations in low-resource settings — do not appear in the document that defines what the world’s antibiotic developers should prioritise.
CARB-X, which I covered in March, has funded neonatal sepsis diagnostics and novel chemistry. It has not issued a call for oral enteric antibiotic development. The AMR Action Fund, with $1 billion in commitments, has invested in late-stage clinical candidates — all IV, all hospital-focused. The entire incentive architecture of antibiotic development is oriented toward critical care, not community medicine.
This is not irrational. Hospital infections kill faster and generate more dramatic data. IV drugs have clearer regulatory pathways. But the result is a pipeline shaped by regulatory convenience rather than disease burden — a pipeline that produces weapons for the 1.27 million AMR deaths in hospitals while ignoring the hundreds of millions of infections that happen outside them.
The vaccine is not a substitute
Typhoid conjugate vaccines are one of global health’s genuine successes. Gavi has supported vaccination of over 130 million children. Pakistan’s mass campaigns have measurably reduced XDR typhoid incidence. But vaccines have structural limitations that make them insufficient as a standalone response to the oral treatment desert.
There is no licensed Shigella vaccine. The most advanced candidate, S4V2, is in Phase 2, with data expected in the second half of 2026. GSK transferred its altSonflex programme to Bharat Biotech, where a Phase 3 design is under development. Shigella vaccine development has been stalled for decades because the organism’s serotype diversity — and the dominance of different species in different geographies — makes a universal vaccine extraordinarily difficult.
And even perfect vaccines cannot treat the patient already infected. The physician in Los Angeles looking at that stool culture needs an antibiotic, not a vaccine. The child in Dhaka with bloody dysentery needs an oral drug that works today, not a vaccine that might be available in 2030.
The azithromycin thread
One drug ties all three disease stories together. Azithromycin was the last oral antibiotic standing for XDR Shigella. It is the last oral option for XDR typhoid. It was part of the recommended dual therapy for gonorrhea. In all three diseases, it is failing — or has already failed.
This is not coincidence. As I documented in The Reckoning, the COVID-19 pandemic drove an estimated 38 million excess azithromycin courses in India alone, accelerating macrolide resistance in enteric bacteria that had nothing to do with COVID. The MORDOR follow-up showed that mass drug administration of azithromycin to children in sub-Saharan Africa produced resistance that persisted 3.5 years after treatment stopped, spreading to communities that never received the drug. And the Toto Bora trial in Kenya revealed the sharpest finding of all: in children carrying mef(A) resistance genes, azithromycin treatment caused 2.72-fold increased rehospitalisation or death. The drug did not just fail. It harmed.
The azithromycin story is the story of a single molecule asked to carry too many burdens — paediatric MDA, COVID panic prescribing, STI prophylaxis, enteric treatment of last resort — and collapsing under the weight of its own overuse. Its failure did not happen in one disease. It happened in all of them at once, because the resistance it generated is not disease-specific. It is bacterial. And bacteria do not respect the boundaries between clinical specialties.
What would need to happen
The oral treatment desert is not a problem that requires new science. The drugs exist. The susceptibility data exists. What is missing is the institutional will to connect them.
A WHO target product profile for oral enteric antibiotics — defining what “good enough” looks like for oral treatment of XDR Shigella and typhoid, including paediatric formulations, stability at tropical temperatures, and activity against known resistance mechanisms.
Regulatory pathways for enteric indication expansion — enabling tebipenem, fosfomycin, and pivmecillinam to be studied for shigellosis and typhoid without requiring each manufacturer to independently fund full Phase 3 programmes. GARDP, which shepherded zoliflodacin from a stalled clinical programme to FDA approval and global access rollout, is the model.
Accelerated Shigella vaccine development — with the urgency currently reserved for pandemic pathogens. 188 million cases and 164,000 deaths per year, with rising XDR prevalence, meets any reasonable threshold for emergency investment.
Genomic surveillance of enteric AMR at the scale of respiratory surveillance — because the blaDHA-1 genotype in Los Angeles was only found because someone sequenced it. Without routine WGS of enteric isolates, novel resistance mechanisms circulate undetected until clinical failure forces the question.
None of this will happen by June 18, when the FDA decides on tebipenem. If approved, tebipenem will be marketed for UTI. Physicians will prescribe it off-label for XDR Shigella and typhoid. Some patients will be saved. There will be no controlled trial data, no paediatric dosing guidance for enteric infections, no pharmacovigilance system tracking off-label outcomes. The oral treatment desert will be navigated by improvisation, not by design.
That is where we are. Three diseases, 200 million cases, and a handful of pills that might work but have never been asked to prove it.
This is Post #31. Sources: CDC MMWR April 2026 (XDR Shigella surveillance), Emergence of novel XDR S. sonnei carrying blaDHA-1 in Los Angeles, Azithromycin-resistant S. Typhi systematic review, First NDM-5 carbapenem-resistant S. Typhi, Spero/GSK tebipenem pipeline. Previously: The Reckoning (Post #29), The Gonorrhea Turnaround (Post #9), The Other Factory (Post #27), The Blueprint (Post #25).