The United States faces a substantially larger estimated burden of antimicrobial resistance (AMR) than earlier national assessments captured. According to the Centers for Disease Control and Prevention (CDC), drug-resistant bacterial and fungal pathogens cause more than 9.3 million infections and 39,000 deaths annually.1 Including Clostridioides difficile, which is not typically drug-resistant but shares important drivers with AMR, brings those figures to approximately 9.7 million infections and 50,800 deaths. However, the CDC explicitly cautions against comparing these estimates with its previous report to determine whether resistance has increased.
This apparent contradiction is central to understanding the agency’s 2026 Antimicrobial Resistance Threats Report, released on October 27.1 The report incorporates broader data sources, revised case definitions, and new estimation methods to characterize infections that earlier assessments did not fully capture. It classifies 19 bacterial and fungal threats as urgent, serious, or concerning and identifies six additional pathogens or pathogen groups on a Watch List, including several whose national burden remains poorly understood.
The assessment raises questions beyond the number of infections and deaths attributed to resistance. Which changes reflect a genuine shift in epidemiology, and which reflect improvements in detection and measurement? How should public health authorities respond when resistance is emerging but its prevalence remains uncertain? What evidence is needed to translate surveillance findings into clinical decisions, prevention strategies, and priorities for antimicrobial development?
The report provides a broader account of established threats while drawing attention to infections that remain difficult to identify, quantify, and monitor.
What the New Numbers Measure and What They Do Not
Although published in 2026, the report’s national burden estimates describe infections and deaths in 2023. They are statistical estimates developed from multiple surveillance systems and other data sources rather than a direct count of every resistant infection diagnosed in the United States.2 Surveillance coverage, diagnostic practices, and methods used to estimate unobserved infections differ substantially among pathogens.
The CDC combined pathogen-specific estimates using probability distributions representing the available evidence and associated uncertainty. The resulting estimate of 9.7 million infections, including C. difficile, has a 95% uncertainty interval of 8.2 million to 11.7 million. The corresponding estimate of 50,800 deaths has an interval of 39,400 to 65,100. The agency also adjusted for overlapping resistance categories to prevent double counting, including removing carbapenem-resistant Enterobacterales infections from the extended-spectrum cephalosporin-resistant Enterobacterales category when calculating the combined burden.
Changes to the underlying case definitions are equally consequential. The CDC expanded its estimates for drug-resistant group A Streptococcus and clindamycin-resistant group B Streptococcus to include noninvasive infections. It also broadened the extended-spectrum cephalosporin-resistant Enterobacterales category to include additional organisms, including Enterobacter species and Klebsiella aerogenes. These revisions incorporate infections and bacterial groups that earlier estimates did not represent in the same way.
The treatment of C. difficile provides another example. Although the organism is not generally considered drug-resistant, antibiotic exposure can disrupt the normal intestinal microbiome, allowing C. difficile to proliferate and cause infection. Antibiotic use and infection prevention are therefore important concerns for both C. difficile disease and AMR. Earlier threat reports included hospitalized C. difficile cases, while the 2026 report expands the estimate to encompass hospitalized and nonhospitalized patients.3 The agency’s death estimate, however, specifically concerns deaths among hospitalized patients.
These changes make the assessment more comprehensive but complicate historical interpretation. Differences between the 2019 and 2026 estimates may reflect revised case definitions, expanded surveillance coverage, different underlying datasets, or changes in modeling methods. They cannot reliably establish how much resistance increased or decreased between the reporting periods.
One important exception is Candida auris, an emerging fungal pathogen that can cause severe infections and spread within healthcare facilities.4 The CDC considers comparisons between its 2019 and 2026 reports appropriate for this organism. Reported C. auris clinical cases increased from 477 in 2019 to 4,523 in 2023, approximately a 9.5-fold rise. These cases involve organisms identified in specimens collected for clinical purposes and do not necessarily represent symptomatic infection in every patient. The CDC also documented increased screening detections, partly reflecting expanded efforts to identify colonized patients.
The rise in reported C. auris clinical cases demonstrates what comparable surveillance can reveal. For most other pathogens, the revised national estimates establish a new assessment of burden rather than a reliable historical trend.
What Has Actually Changed in the Threat Landscape?
The CDC’s revised classifications provide another perspective on the evolution of resistance. Individual surveillance programs can identify shifts in susceptibility even when national burden estimates cannot be compared across reports. Meanwhile, the agency can revise a threat’s classification as evidence changes concerning its clinical consequences, transmission, preventability, or available treatments.
The CDC considers seven factors when assessing threats, including current and projected burden, clinical and economic consequences, transmissibility, preventability, and treatment options. The urgent, serious, and concerning categories therefore represent multidimensional assessments rather than rankings based solely on infection or death counts.
One of the most consequential additions in 2026 is drug-resistant Neisseria meningitidis, newly classified as a serious threat. The organism causes meningococcal disease, which can produce rapidly progressing meningitis and bloodstream infections. Although relatively uncommon in the United States, meningococcal disease requires prompt antibiotic treatment and preventive antibiotic administration to close contacts because severe illness can develop quickly.
The CDC estimates that penicillin-resistant meningococcal cases increased from 19 in 2019 to 61 in 2023, while ciprofloxacin-resistant cases increased from 16 to 35. The estimated proportion of cases resistant to penicillin rose from 5.1% to 14.0%, while ciprofloxacin resistance increased from 4.3% to 8.0%.5 The CDC notes overlapping credible intervals for several annual comparisons, so individual year-to-year differences should not be interpreted as statistically established increases. Nevertheless, the broader pattern raises concern about the clinical utility of these antibiotics.
Penicillin resistance can complicate decisions about narrowing treatment once susceptibility results become available. Ciprofloxacin resistance presents a different concern because the antibiotic is used for post-exposure prophylaxis among close contacts of people with meningococcal disease. Resistant strains may reduce the effectiveness of that preventive intervention, potentially allowing additional cases to develop.
In February 2024, Isha Berry and colleagues at the CDC and collaborating public health institutions published implementation guidance in the Morbidity and Mortality Weekly Report addressing ciprofloxacin resistance.6 The guidance recommends that health departments preferentially consider alternatives to ciprofloxacin when two conditions are met within a local catchment area during a rolling 12-month period: at least two invasive meningococcal disease cases caused by ciprofloxacin-resistant strains, and resistance in at least 20% of all reported invasive meningococcal cases. Recommended alternatives include rifampin, ceftriaxone, and azithromycin. The guidance responds to the risk of prophylaxis failure, not evidence that resistant meningococci had already caused widespread prophylaxis failures.
These thresholds demonstrate how surveillance can inform geographically targeted preventive measures. A pathogen does not need to account for a large share of national infections to warrant attention when resistance threatens an intervention used to prevent further disease.
The reclassification of C. difficile illustrates a different type of change. Previously categorized as urgent, the pathogen is now classified as serious, reflecting declines in cases and progress in prevention, particularly within healthcare settings. The CDC nevertheless estimates that C. difficile caused 412,504 infections among hospitalized and nonhospitalized patients in 2023, along with 11,339 deaths among hospitalized patients.3 It also estimates more than $6.4 billion in attributable healthcare costs.
The CDC attributes the revised classification partly to improvements in infection prevention, but the epidemiological trajectory is not a continuous decline. Estimated infections decreased between 2018 and 2020 before increasing somewhat thereafter, and changes in diagnostic practices complicate interpretation of the trend. Community-associated cases also now exceed healthcare-associated cases, reinforcing the importance of prevention and appropriate antibiotic use beyond hospitals.
The two examples show why changes in classification require context. Meningococcal resistance has direct implications for antibiotic selection and prophylaxis; the revised C. difficile designation reflects progress against an established threat whose remaining burden is substantial.
Emerging Threats That Surveillance Cannot Yet Fully Characterize
The CDC’s six-category Watch List identifies potential threats whose full burden in the United States remains uncertain. Unlike the urgent, serious, and concerning categories, inclusion does not require a comprehensive national estimate of infections or deaths. Three entries were added in 2026: drug-resistant dermatophytes, drug-resistant Haemophilus influenzae, and nontuberculous mycobacteria (NTM). The NTM entry reflects the burden and treatment challenges posed by a diverse group of environmental bacteria, many of which are difficult to treat because of resistance; it does not imply that every NTM infection involves newly acquired resistance.7
Drug-resistant dermatophytes provide a clear example of resistance that may escape routine surveillance. Dermatophytes are fungi that infect the skin, hair, and nails, causing conditions commonly known as ringworm or tinea. Most infections respond to topical antifungal medications, but resistance to treatments such as terbinafine has emerged in organisms including Trichophyton indotineae and Trichophyton rubrum. Resistant infections can be more extensive, require prolonged or alternative therapy, and recur after treatment.8
Determining how frequently these infections occur is difficult. Dermatophyte infections may be misdiagnosed or go unrecognized, and routine clinical assessment does not necessarily identify the species responsible or establish its susceptibility to treatment. Specialized identification and antifungal susceptibility testing are not widely available, and dermatophyte infections are not nationally reportable in the United States. The CDC consequently lacks the data needed to establish their national burden reliably.
The agency identifies inappropriate antifungal prescribing, misuse of over-the-counter antifungals, and antifungal–corticosteroid combination products as factors contributing to resistance. It is working with state and local partners to improve detection, including expanding laboratory capacity for identification and antifungal susceptibility testing through the Antimicrobial Resistance Laboratory Network. Without reliable identification of resistant infections, it is difficult to determine whether individual cases reflect isolated treatment failures or wider circulation of resistant strains.
NTM present a different epidemiological challenge. These bacteria occur naturally in soil, dust, and water, including building plumbing systems. Most healthy individuals exposed to NTM do not develop disease, but the organisms can cause pulmonary, bloodstream, lymph node, skin, and soft-tissue infections. Pulmonary disease is more common among people with underlying lung disease or weakened immune systems.7
Many NTM are resistant to multiple antibiotics, and susceptibility varies considerably among species. Treatment often requires prolonged multidrug regimens tailored to the infecting organism, disease site, and susceptibility profile. Some pulmonary infections require treatment for more than a year, while recurrence, relapse, and the occasional need for surgery further complicate management.
The CDC identifies environmental exposure, particularly to water, as an important source of NTM infections, although the source of an individual infection may be difficult to establish. Healthcare-associated infections can also occur when contaminated water or medical equipment introduces organisms into susceptible tissues. Because NTM are widespread in the environment, identifying an organism does not necessarily establish that it caused disease requiring treatment. Estimating burden therefore depends on clinical interpretation alongside microbiological identification.
Drug-resistant H. influenzae, the third new Watch List addition, illustrates the importance of monitoring susceptibility before resistance becomes widespread enough to compromise established therapies.9 The bacterium can colonize the upper respiratory tract without causing illness, but it can also cause ear infections, pneumonia, bloodstream infections, and meningitis. Resistance to ampicillin, amoxicillin, and penicillin is common, while resistance to amoxicillin-clavulanate remains uncommon but is increasing. Ceftriaxone resistance is rare in the United States, although resistant infections have been reported internationally.
The CDC monitors invasive H. influenzae disease through Active Bacterial Core surveillance and national disease reporting. It also receives isolates and uses whole-genome sequencing to identify mutations associated with resistance. These capabilities provide an established basis for monitoring invasive disease, but they do not necessarily characterize resistance across the broader spectrum of respiratory and other noninvasive infections.
The Challenge of Surveillance Beyond Hospitals
The expanded scope of the 2026 report raises questions about how resistance data are collected and integrated across healthcare settings. Hospitals remain essential surveillance environments because they care for patients with severe infections and provide opportunities for microbiological testing, infection monitoring, and investigation of transmission. However, the infections represented in hospital data sets differ from those diagnosed and managed in outpatient clinics, long-term care facilities, and communities.
A surveillance system focused on invasive bacterial infections can provide valuable information about severe disease, mortality, and resistance among organisms recovered from normally sterile body sites. It may provide less information about noninvasive infections treated outside hospitals, where cultures and susceptibility testing are not always obtained. Similarly, systems designed to monitor healthcare-associated infections cannot necessarily establish the burden of infections acquired or treated elsewhere.
For example, if a clinic increasingly cultures patients with recurrent infections or prior treatment failure, its measured proportion of resistant isolates could rise even without a comparable increase among all patients with that infection. This is a sampling effect rather than proof of a population-wide trend. Interpreting susceptibility data therefore requires knowing who was tested, why testing occurred, and whether those practices changed.
The CDC’s inclusion of noninvasive streptococcal infections and nonhospitalized C. difficile cases reflects the importance of accounting for differences among settings. Nevertheless, broader coverage introduces methodological challenges. Hospital, outpatient, and public health laboratory datasets may represent different populations and use different diagnostic criteria, sampling practices, and reporting procedures. Combining them requires attention to representativeness, overlapping cases, and statistical uncertainty.1,2
An increase in detected resistance may reflect a genuine epidemiological change, but expanded testing, improved identification, or broader reporting can also increase recognized cases. More representative surveillance consequently requires both broader coverage and enough methodological consistency to distinguish changes in disease from changes in detection. Clinical data, microbiological and genomic characterization, and epidemiological investigation each contribute different evidence to that assessment.
Turning Surveillance Findings Into Public Health Decisions
The practical value of surveillance depends on the decisions it can support. For some pathogens, the priority is detecting changes in susceptibility that affect established treatments or preventive interventions. For others, the immediate challenge is identifying a resistant organism reliably enough to determine whether further investigation is warranted.
The meningococcal prophylaxis guidance provides an example of surveillance linked to a defined action threshold. Rather than relying solely on national resistance prevalence, the CDC recommends that health departments consider both the number and proportion of resistant cases within a geographic area. This allows public health recommendations to respond to local resistance patterns without requiring a nationwide change in prophylaxis practice.
Identifying a resistance signal and determining whether it justifies intervention are separate tasks. A resistant isolate may warrant investigation, but decisions to change clinical or public health practice depend on additional evidence about the frequency, distribution, and consequences of resistance. The evidence required differs according to the intervention under consideration.
For emerging pathogens, waiting until a comprehensive national burden estimate becomes available may delay investigations needed to establish whether resistance is spreading. The Watch List acknowledges this problem by identifying organisms that warrant monitoring despite incomplete evidence of population-level impact. The appropriate response depends on the uncertainty involved: laboratory identification and susceptibility testing, epidemiological investigation, or broader surveillance of infections that existing systems capture incompletely.
The CDC has also introduced an interactive online format intended to support more frequent updates to its threat assessments. This approach may communicate newly available evidence without waiting for another comprehensive report. Its effectiveness will depend on the timeliness, representativeness, and quality of the underlying data, as well as clear documentation of methodological changes.
These distinctions matter for antimicrobial research and development. National burden estimates help establish the scale of unmet medical need, but substantial disease burden alone does not establish the rationale for developing a particular antimicrobial. Developers must also understand which resistance phenotypes are clinically important, whether affected patients can be identified and recruited for studies, which outcomes can demonstrate benefit, and how a proposed intervention would improve upon available treatment options.
The meningococcal example illustrates how emerging resistance can affect the utility of an established antibiotic even when resistant infections remain comparatively uncommon. Dermatophytes present a different development challenge: limited susceptibility testing complicates efforts to estimate resistance prevalence, identify patients who might benefit from alternative antifungals, and establish the clinical evidence needed to evaluate those treatments.
For pathogens with uncertain burden, improving surveillance can contribute to development feasibility as well as public health understanding. Better microbiological characterization can define target populations and relevant resistance phenotypes, while clinical and epidemiological data can clarify outcomes and settings for evaluating new interventions. These requirements are especially important when resistant infections are uncommon, geographically concentrated, or difficult to distinguish from susceptible infections using routine diagnostics.
The CDC’s classifications provide a starting point for identifying unmet needs, but they cannot substitute for the detailed evidence needed to design and evaluate individual diagnostic, preventive, or therapeutic interventions.
Measuring Emerging Threats Before They Become Established
The CDC’s 2026 report expands the national picture of AMR while clarifying the limits of historical comparisons. Its revised estimates capture infections that earlier assessments did not fully represent, while pathogen-specific surveillance identifies resistance changes that already influence public health recommendations.
The Watch List identifies organisms warranting attention before their national burden can be estimated confidently. Waiting for comprehensive data could delay recognition of important resistance problems, but treating every emerging signal as evidence of a widespread threat could misdirect limited resources. The challenge is to identify emerging resistance early enough to change clinical practice and direct research, without confusing incomplete surveillance with evidence of a large or growing burden.
