In 2019, New Delhi metallo-β-lactamase (NDM) was a relatively uncommon finding in American hospitals. By 2023, NDM-producing carbapenem-resistant Enterobacterales (NDM-CRE) had surged 461%. In New York City, NDM surpassed KPC — the previously dominant carbapenemase — as the most frequently reported resistance mechanism in 2024.
This is not a gradual trend. It is an inflection point. And it demands attention, because NDM is not just another resistance gene. It is strategically different from what came before — and our therapeutic toolkit was not built for it.
What Makes NDM Different
To understand why NDM is an emergency, you need to understand the two main families of carbapenemases — the enzymes that destroy our last-resort antibiotics.
KPC (Klebsiella pneumoniae carbapenemase) is a serine β-lactamase. It breaks down carbapenems, but it can be inhibited by newer β-lactamase inhibitors like avibactam, vaborbactam, and relebactam. This is why the drugs we developed in the 2010s — ceftazidime-avibactam, meropenem-vaborbactam, imipenem-relebactam — work against KPC-producing bacteria. For a decade, KPC dominated the American CRE landscape, and we had an answer.
NDM is a metallo-β-lactamase (MBL). It uses zinc ions to hydrolyze β-lactams through a fundamentally different mechanism. And here is the critical difference: none of the newer β-lactamase inhibitors work against it. Avibactam, vaborbactam, relebactam — all powerless against NDM. The drugs we built to fight KPC are irrelevant against NDM.
| Drug | vs KPC | vs NDM |
|---|---|---|
| Ceftazidime-avibactam | ✓ | ✗ |
| Meropenem-vaborbactam | ✓ | ✗ |
| Imipenem-relebactam | ✓ | ✗ |
| Emblaveo (aztreonam-avibactam) | ✓ | ✓ |
| Cefiderocol | ✓ | ✓ |
| Zaynich (cefepime-zidebactam) FDA REVIEW | ✓ | ✓ |
Think of it this way. We spent a decade building an arsenal specifically optimized for KPC. Now the enemy has shifted to a fundamentally different mechanism — one that makes most of that arsenal obsolete. The bacteria did not just evolve resistance to our drugs. They evolved around the entire class of inhibitors we developed.
The Numbers
The CDC reported in September 2025 that carbapenemase-producing CRE (CP-CRE) increased 69% between 2019 and 2023 across 29 US states. But that headline number obscures the real story. KPC cases increased modestly. The explosion was driven almost entirely by NDM.
New York City tells the story most vividly. NDM-positive CRE cases rose from 58 in 2019 to 388 in 2024 — a nearly seven-fold increase. Age-adjusted incidence went from less than 1 per 100,000 residents to 3.8 per 100,000. And crucially, the CDC data suggest this is no longer confined to healthcare facilities. The increase in NDM-CRE among patients without known long-term care facility exposure points to community transmission — meaning NDM is spreading beyond the hospital walls where we have some capacity to contain it.
By 2023, NDM was found in 27% of carbapenem-resistant E. coli and 24% of carbapenem-resistant Klebsiella species nationwide. NDM had surpassed KPC in E. coli specifically. For Klebsiella, KPC still dominates — but the trajectory is clear.
A Global Phenomenon
The US surge is part of a worldwide pattern. NDM is not staying where it started.
Mexico: A 2026 study of pediatric CRE at a major hospital found NDM in 52% of isolates. The mortality rate was 47.5%. Children, dying of infections that resist our strongest antibiotics. And it is not confined to hospitals — blaNDM-5 was detected in community air samples and wastewater in Frontiers in Microbiology (March 2026).
Virginia, USA: NDM-7, a variant previously associated primarily with South Asia, has been found in patients with no travel history whatsoever. Genomic analysis reveals NDM-7 is carried exclusively on IncX3 plasmids across 12 different bacterial species. While NDM-1 and NDM-5 cases still correlate with international travel, NDM-7 appears to be transmitting domestically. In Virginia, NDM rose from 6% to 30% of all carbapenemase-producing organism isolates.
Canary Islands, Spain: An outbreak of NDM-14-producing K. pneumoniae ST147 spread across 99 patients in 4 hospitals on 3 islands, traced to a link with Morocco (CDC Emerging Infectious Diseases, January 2026).
Sweden: After more than a decade of negative surveillance, NDM-5 Enterobacterales were detected for the first time in wastewater treatment plant effluent in Kristianstad (January 2026). Some strains carried dual carbapenemases — NDM-5 plus OXA-181. Even one of the world’s most careful countries is finding NDM in its water.
Companion animals: A CDC Emerging Infectious Diseases study (March 2026) found genetically similar NDM-5 CP-CRE strains in companion animals and humans in the United States. Of 246 companion animal CP-CRE sequences, 69% clustered with human clinical isolates. The dominant allele: blaNDM-5. The resistance network extends into our homes, through our pets. Direction of transmission remains unknown.
The Evolutionary Arms Race Intensifies
Perhaps the most alarming development is what happens when bacteria face both KPC-targeting drugs and NDM pressure simultaneously.
KPC-NDM co-production — bacteria carrying both carbapenemase types on separate plasmids — is emerging globally. Argentina’s RECAPT-AR national surveillance study found 3.4% of carbapenemase-producing Enterobacterales co-harbored both NDM and KPC. In a Thessaloniki intensive care unit study (2025), 50% of isolates carried both blaKPC and blaNDM. In China, the ST11-KL64 K. pneumoniae lineage has been identified as a vehicle for KPC-NDM co-production.
These dual producers are resistant to everything. Ceftazidime-avibactam fails (NDM hydrolyzes aztreonam independently, and avibactam cannot inhibit NDM). Meropenem-vaborbactam fails. Imipenem-relebactam fails. We are approaching pan-resistance — infections that no approved drug can reliably treat.
This co-production is not random. It appears to be an evolutionary response to the selective pressure of ceftazidime-avibactam itself. As hospitals deploy KPC-targeting drugs, bacteria that also carry NDM gain a survival advantage. We are selecting for the worst possible combination.
And there is a subtler threat. Researchers have identified “silent NDM” — K. pneumoniae strains carrying blaNDM-1 with a promoter deletion that renders the gene phenotypically silent. These bacteria appear susceptible on standard laboratory testing. They pass the AST. They look treatable. But the gene is there, ready to reactivate under antibiotic pressure or transfer to another organism. A hidden reservoir, invisible to our surveillance.
The Counterattack
We are not defenseless. But our options against NDM are narrower than most clinicians realize, and the timeline is tight.
Emblaveo (aztreonam-avibactam) — Approved February 2025
AbbVie’s Emblaveo is the first fixed-dose monobactam/β-lactamase inhibitor combination. The logic is elegant: aztreonam is the one β-lactam that NDM cannot hydrolyze (monobactams are structurally immune to metallo-β-lactamases). But NDM-producing bacteria usually co-produce other β-lactamases that can destroy aztreonam. Adding avibactam blocks those other enzymes, restoring aztreonam’s activity. It is currently approved for complicated intra-abdominal infections in the US, with broader indications (pneumonia, UTIs) approved in Europe.
The limitation: Emblaveo’s approval was based on relatively modest clinical data (the REVISIT and ASSEMBLE trials), and the drug is only available as IV therapy requiring hospitalization. Real-world outcome data for NDM-specific infections is still being collected.
Cefiderocol — Approved 2019
Cefiderocol uses a siderophore mechanism — it mimics iron transport molecules to sneak through bacterial outer membranes, then attacks penicillin-binding proteins from inside. It retains activity against NDM producers in most cases. But it is not universally reliable: some NDM-producing strains show reduced susceptibility, and a concerning signal from the CREDIBLE-CR trial showed higher all-cause mortality in the cefiderocol arm (34%) versus best available therapy (18%) in a subgroup of Acinetobacter patients. It remains a critical tool, but not a silver bullet.
Zaynich (cefepime-zidebactam) — FDA Priority Review
Wockhardt’s Zaynich may be the most important anti-NDM drug in the pipeline. Zidebactam is a first-in-class β-lactam enhancer that works through dual mechanisms: high-affinity binding to PBP2 (directly blocking cell wall synthesis) and β-lactamase inhibition (protecting cefepime). This dual action gives it activity against NDM producers that other combinations miss.
The Phase 3 data is striking: 96.8% clinical cure rate, superior to meropenem (89.0% vs 68.4% combined clinical-microbiological endpoint). The NDA was accepted in December 2025 with priority review. Over 100 patients have been treated through compassionate use globally, including a pediatric cancer patient at St. Jude’s with NDM-producing E. coli who achieved full recovery after 41 days of treatment.
If approved — likely mid-2026 — Zaynich would significantly strengthen our anti-NDM arsenal. Wockhardt is also developing WCK 6777 (ertapenem-zidebactam), a once-daily outpatient formulation that could extend zidebactam’s reach beyond the hospital.
Imipenem/cilastatin/funobactam (XNW4107) — Phase 3 Complete
Evopoint’s funobactam-based combination has completed Phase 3 trials for both complicated UTI and hospital-acquired pneumonia. As a diazabicyclooctane (DBO) β-lactamase inhibitor, it covers Class A, C, and D β-lactamases including problematic OXA-23 and OXA-24 variants. Activity against NDM specifically is more limited than Emblaveo or Zaynich, but it adds breadth against co-produced enzymes. Pre-NDA filing is underway in China.
What Must Change
The NDM emergency exposes three systemic failures.
First, diagnostic lag. Standard antimicrobial susceptibility testing tells you what a bacterium is resistant to, but not why. Carbapenemase identification — distinguishing KPC from NDM — requires molecular testing that many clinical labs still do not perform routinely. As the CDC noted, limited carbapenemase testing delays targeted treatment and means the true burden of NDM is almost certainly underestimated. When you do not know you are facing NDM, you reach for ceftazidime-avibactam — and it fails. The hidden cost of that diagnostic gap is measured in lives.
And standard AST has an even deeper blind spot. “Silent NDM” strains — phenotypically susceptible, genotypically resistant — pass routine testing entirely. Only molecular methods can catch them. We need genotypic and phenotypic testing running in parallel, and we need it fast.
Second, the One Health gap. NDM is in hospital patients, in pets, in wastewater, in community air samples. It crosses species boundaries on mobile plasmids with disturbing ease. Yet our surveillance systems still operate in silos — human health, veterinary health, and environmental monitoring rarely communicate in real time. The CDC’s finding that 69% of companion animal CP-CRE sequences cluster with human isolates should be a wake-up call. This is one interconnected ecosystem of resistance, and we need to monitor it as such.
Third, the pipeline gap. We have exactly three β-lactam-based options for NDM: Emblaveo (approved for a single indication), cefiderocol (with concerning mortality signals in some populations), and Zaynich (not yet approved). Against 461% growth. Against KPC-NDM co-producers approaching pan-resistance. Against a gene that is spreading into communities, pets, and wastewater. Three drugs is not enough. We need the PASTEUR Act, now in its fourth Congressional introduction, to finally receive a floor vote. We need the UK subscription model — which begins new contracts this April — to prove that pull incentives can sustain antibiotic development. And we need the pipeline to look beyond β-lactam/β-lactamase inhibitor combinations toward fundamentally new mechanisms, like the novel classes I covered in The New Arsenal.
The Race We Cannot Afford to Lose
NDM is not just another resistance mechanism. It represents an evolutionary strategy that renders an entire therapeutic paradigm obsolete. The serine β-lactamase inhibitors that defined the last decade of antibiotic development are powerless against it. And NDM is not waiting — it is spreading through hospitals, communities, pets, and waterways, on promiscuous plasmids that jump between species.
The pediatric ward in Mexico where 47.5% of children with CRE died. The patient in Virginia who never traveled abroad but carries NDM-7. The dog in the CDC study whose E. coli matches a hospitalized human’s strain. The Swedish wastewater plant finding NDM for the first time after a decade of clean results. These are not isolated data points. They are the leading edge of a wave.
We have drugs that work against NDM — some approved, one nearly there. We have the science to develop more. What we lack is the speed, the funding, the diagnostic infrastructure, and the coordinated surveillance to deploy these tools before the wave overwhelms us.
The bacteria are not waiting for Congressional votes or market incentives. They are doubling, spreading, and evolving. The NDM emergency is here. The question is whether we act like it.
Sources: CDC Annals of Internal Medicine (Sep 2025); MMWR (Jun 2025); CDC Emerging Infectious Diseases (Jan & Mar 2026); BMC Infectious Diseases Mexico pediatric CRE (2026); Frontiers in Microbiology community NDM plasmids (Mar 2026); RECAPT-AR Argentina national surveillance (2024); MDPI Thessaloniki ICU dual KPC/NDM (2025); AbbVie Emblaveo FDA approval (Feb 2025); Wockhardt Zaynich NDA/Phase 3 data (2025); PMC Swedish wastewater NDM-5 (Jan 2026).