CDC’s 2026 Antimicrobial Resistance Threats Report puts the annual US burden from its listed bacteria and fungi at an estimated 9.7 million infections and 50,800 deaths. Those figures mainly describe 2023, rather than infections counted in 2026. They also should not be read as a straightforward increase from the headline totals in an older report: CDC changed the methods and the scope of the estimates. CDC report overview
There is another important boundary inside the total. The report includes Clostridioides difficile, or C. diff, even though its infections are not typically drug-resistant. Understanding what enters the count is more useful than treating the headline as a single measure of how much resistance has risen.
The report date and the data year do different jobs
The new report brings together burden estimates for 19 urgent, serious and concerning threats. CDC describes the overall calculation as an aggregation of 2023 estimates, using different data sources and estimation approaches for different pathogens. Its separate Watch List identifies potential threats whose full US burden is not yet understood. A Watch List entry is not an extra measured contribution to the headline total. Threats, categories and methods
Even the 2023 label needs its footnotes. The burden table reports tuberculosis disease cases for 2023 but tuberculosis deaths from 2021, reflecting the reporting time needed to include treatment outcomes. A release year, a case year and an outcome-reporting year can therefore appear in the same publication without being interchangeable. US burden table, both pages (PDF)
For someone using these figures in a briefing or service plan, the useful citation is the named estimate, population and data period—not simply “CDC figures for 2026”.
Why the new and old totals cannot establish a trend
CDC says its methodologies and data sources changed substantially between previous reports and this update for all pathogens except C. auris. The agency therefore advises against direct comparisons between reports. This is more than a warning about rounding.
Some definitions now cover more disease. The updated group A and group B streptococcal estimates include non-invasive infections as well as invasive disease. The C. diff estimate includes hospitalized and non-hospitalized cases, whereas previous reports included hospitalized cases only. The expanded cephalosporin-resistant Enterobacterales category also includes additional organisms. Changes since the previous report
A larger total can reflect a wider counting frame or a different estimation method. Establishing a change in burden requires comparable observations; subtracting two unlike report totals cannot supply that comparison.
C. diff explains why the title is broader than resistant infections alone
Antibiotics can disrupt the normal balance of the gut microbiome and allow C. diff to cause infection. CDC includes the organism because antibiotic use and the spread of germs are important to both C. diff prevention and resistance prevention. That connection does not make every C. diff case a resistant infection. CDC’s C. diff explanation
The agency also moved C. diff from the urgent category in earlier reports to serious in this update. CDC attributes that change partly to successful prevention, especially in healthcare settings, while retaining a substantial estimated burden. “Serious” is still a category requiring sustained action; it does not mean the disease has become unimportant.
Threat categories are judgments about several factors, including clinical impact, spread, preventability and available treatments. They are not a league table ordered only by the number of infections.
The total carries uncertainty—and avoids some double counting
The published interval around 9.7 million infections is 8.2 million to 11.7 million. For 50,800 deaths, it is 39,400 to 65,100. The table describes these as confidence or credible intervals, depending on the underlying estimate. Burden estimates and uncertainty (PDF)
CDC’s aggregation combines probability distributions rather than merely adding every displayed row and its uncertainty bounds. Its methods assume the included pathogen distributions are independent. The accompanying documents also explain overlap handling: CRE is removed from the broader ESC-RE category when combining estimates, and C. auris is not added separately because hospitalized drug-resistant C. auris infections are already included within drug-resistant Candida. Aggregation methods (PDF)
Those details make the report a useful national reference for understanding the scale and composition of the problem. They also set the boundaries for using it: preserve the data period, retain uncertainty, check the pathogen definition and use comparable evidence before describing a rise or fall.