The professional society that writes the treatment playbook for America's hardest-to-treat bacterial infections has issued a new version, and the timing is not coincidental. The Infectious Diseases Society of America published its updated guidance on antimicrobial-resistant Gram-negative infections in Clinical Infectious Diseases, replacing earlier versions that hospitals have relied on.
The update matters because the bacteria moved first. Federal laboratory surveillance has documented a sharp rise in a resistance gene that disables not only last-resort antibiotics but also several newer drugs specifically developed to combat resistance, and clinicians have been making treatment decisions in that gap.
For most people going about a normal week, this is not a personal threat. The risk is concentrated in hospitals, particularly among patients in intensive care with ventilators, catheters, or long-term admissions. But the guidance shapes what a hospital can offer a family member who ends up in that situation and how quickly the right drug can be started.
Newer Drugs Only Help If the Laboratory Names the Enzyme
Carbapenem-resistant Enterobacterales, or CRE, are gut bacteria, including E. coli and Klebsiella pneumoniae, that have become resistant to carbapenems, a class of antibiotics normally held in reserve. CDC classifies CRE as an urgent threat.
Resistance in these organisms usually comes from enzymes called carbapenemases, and which enzyme a patient's bacteria carries determines which drugs will work. For decades, the dominant enzyme in the United States was KPC. That has been changing. A CDC analysis published in Emerging Infectious Diseases found that among 1,288 carbapenemase-producing CRE collected through the agency's multi-site surveillance program, the share carrying the blaNDM gene rose from 5.4 percent in 2016 to 39.8 percent in 2023, while the share carrying blaKPC fell from 92.8 percent to 53.0 percent.
New Delhi metallo beta lactamase, the enzyme NDM stands for, is the harder problem. Study authors described the increase as "alarming," as reported by CIDRAP, because NDM-producing CRE resist not only older drugs but also several newer beta-lactam combination antibiotics developed specifically for CRE infections.
The updated IDSA document addresses that directly. According to a summary of changes, the CRE section now incorporates aztreonam-avibactam for NDM-producing infections and adds material on gepotidacin, pivmecillinam, and intravenous fosfomycin. The panel also updated susceptibility breakpoint tables to current standards and revised the laboratory criteria for presuming extended-spectrum beta-lactamase production.
There is a catch built into all of it. A drug that works against NDM only helps if the laboratory identifies NDM, and many clinical laboratories do not routinely perform carbapenemase testing. The consequence is that a patient started on a seemingly reasonable antibiotic that the organism would never respond to. IDSA continues to recommend involving an infectious disease specialist in the management of these infections.
Hospitalized Patients Carry Nearly All of the Risk
CRE infections cause pneumonia, bloodstream infections, urinary tract infections, and wound infections. CDC estimated roughly 13,387 CRE infections among hospitalized patients in 2022. These are not community illnesses in any ordinary sense.
The people at highest risk share a recognizable profile: extended hospital stays, intensive care admission, mechanical ventilation, central lines or urinary catheters, recent broad-spectrum antibiotic exposure, and transfers between health care facilities, including long-term care. Patients with weakened immune systems face compounded risk.
Geography matters too. A report from New York City health officials documented NDM-positive CRE cases rising from 58 in 2019 to 388 in 2024, at which point NDM surpassed KPC as the city's most common carbapenemase. That local shift changes the correct empiric antibiotic choice for a patient in a New York hospital in a way it may not for a patient elsewhere, which is why the guidance stresses knowing local resistance patterns.
New work presented at ASM Microbe 2026 suggests the national increase is not random. Researchers sequenced more than 8,000 NDM-CRE isolates collected through the Antimicrobial Resistance Laboratory Network and found the rise driven largely by a limited set of high-risk bacterial lineages, chiefly Klebsiella pneumoniae and E. coli, including a newly expanding Klebsiella strain designated ST45. Richard Stanton, PhD, a CDC bioinformatician and corresponding author, said in an ASM statement that "surveillance efforts and research into new treatments and diagnostics should prioritize these strains."
That is a finding about strategy, not a treatment recommendation, and it has not changed clinical guidance on its own.
Stewardship Decisions Reach Ordinary Prescriptions Too
The hospital story has a household counterpart, and it is less dramatic but more widely relevant. Antimicrobial stewardship is the practice of using antibiotics only when they will help, in the right dose, for the right duration. Every unnecessary course of action imposes selection pressure that favors resistant organisms.
Practical steps exist for ordinary patients. Antibiotics do not treat colds, most sore throats, or flu. Leftover antibiotics should not be saved or shared, and a prescribed course should be completed as directed rather than stopped when symptoms improve, unless a clinician says otherwise. Anyone hospitalized or with a family member in the hospital can ask whether cultures were taken, whether the antibiotic can be narrowed once results return, and how long the course should run. Those are reasonable questions, and stewardship teams welcome them.
Cost and access are real barriers to these infections. Newer agents active against resistant organisms are expensive and often restricted to inpatient use, and molecular testing that identifies the specific enzyme is not universally available. Patients facing coverage problems can ask about prior authorization, appeals, and manufacturer patient assistance programs.
Several questions remain unresolved. Investigators have not established why NDM expanded so quickly, whether patient outcomes differ meaningfully between NDM and KPC infections, or how widely the ST45 lineage will spread. IDSA has said it intends to update the guidance periodically, and CDC surveillance data will continue to refresh through the laboratory network. MedicalDaily will follow both.
Key Questions Answered
What exactly changed? IDSA published an updated guidance document for the treatment of antimicrobial-resistant Gram-negative infections, replacing earlier versions. It adds newer agents, including aztreonam-avibactam, for NDM-producing infections and updates laboratory breakpoints and criteria.
Is this a threat to healthy people at home? Not primarily. CRE infections occur overwhelmingly in hospitals and long-term care settings among patients with catheters, ventilators, long stays, or weakened immune systems.
What does NDM mean? New Delhi metallo-beta-lactamase is an enzyme that breaks down most available antibiotics, including several newer drugs built specifically to defeat other resistance enzymes.
Why can't doctors just use the newest antibiotic? The drug has to match the resistance mechanism. Aztreonam avibactam works against NDM, but only if the laboratory identifies NDM, and many clinical labs do not routinely test for carbapenemases.
Does taking fewer antibiotics actually help? Yes, at the population level. Every unnecessary course favors the survival of resistant organisms. Antibiotics do not treat colds, flu, or most sore throats.
What can a patient or family member ask in the hospital? Whether cultures were collected, whether the antibiotic can be narrowed once results come back, how long the course should last, and whether an infectious diseases specialist is involved.
Are treatments improving? Some. Aztreonam avibactam gives clinicians an option against NDM that did not exist a few years ago, and other agents were added to the updated guidance. Access and testing remain uneven.
