Resistant hospital infections killed nearly 40% within a month, Asian network finds

A large prospective study across Asian hospitals has put hard numbers on how lethal drug-resistant health-care-associated infections have become: among patients with such infections, more than a third were dead within 28 days [s1]. The findings, from the Asian Surveillance Network ACORN-HAI and published in The Lancet Infectious Diseases on 21 August 2026, move the estimate of antimicrobial resistance (AMR) burden off cross-sectional laboratory data and onto patients followed in real time [s1].

The study enrolled patients of any age with microbiologically confirmed ventilator-associated pneumonia, hospital-acquired bloodstream infection, or health-care-associated bloodstream infection from 41 hospitals across 19 Asian countries and regions, and followed each for 28 days from the onset of infection [s1]. Between 1 September 2022 and 28 February 2025, 10,111 patients were enrolled, of whom 9,496 were included in the final analysis, contributing 9,642 infection episodes — 6,597 (68.4%) bloodstream infections and 3,045 (31.6%) ventilator-associated pneumonia [s1].

How much resistance, and how deadly

Resistance was not the exception but the rule. Of the 9,642 infection episodes, 7,102 — 73.7% — were associated with AMR bacteria, and Gram-negative organisms predominated, accounting for 7,599 (78.8%) of episodes [s1]. That Gram-negative dominance matters clinically, because it is exactly the class of bacteria for which the antibiotic pipeline is thinnest and last-line options are most exhausted.

The mortality that followed those infections was severe. Crude 28-day mortality was 37.7% among the 7,102 AMR infection episodes, rising to 40.6% among the 4,683 episodes involving multidrug-resistant bacteria [s1]. The worst outcomes clustered around two carbapenem-resistant groups: carbapenem- resistant Acinetobacter species carried a crude 28-day mortality of 51.3%, and carbapenem-resistant Enterobacterales 48.4% — meaning that roughly half of patients with those infections died within a month [s1].

Separating the infection from the illness

Crude mortality alone overstates what resistance itself does, because the patients who develop these infections are often already gravely ill. To isolate the toll attributable to resistance, the authors estimated AMR-attributable mortality — the excess death traceable to the resistance rather than to the underlying condition. That attributable mortality was highest in ventilator-associated pneumonia, at 16.9% (95% CI 13.0-20.9) [s1].

Two patterns in the attributable estimates stand out. By age, the burden fell hardest not only on the oldest but on the young and working-aged: AMR-attributable mortality reached 11.7% (95% CI 1.6-21.8) in those aged 5-14 years and 11.2% (95% CI 7.2-15.2) in those aged 15-49 [s1]. And by economic setting, it was highest in lower-middle-income countries, at 10.7% (95% CI 7.0-14.4) — the health systems least equipped with the isolation capacity, newer antibiotics and intensive-care resources that blunt these infections [s1]. By pathogen, the same two carbapenem-resistant groups that drove crude mortality also carried the highest attributable mortality — 19.4% (95% CI 14.1-24.7) for resistant Acinetobacter and 16.2% (95% CI 12.8-19.6) for resistant Enterobacterales — and the poorest quality-of-life outcomes among survivors [s1].

Running out of drugs

The treatment data show clinicians already pushed to the end of the shelf. Most carbapenem-resistant Gram-negative infections were treated with carbapenems (57.8%) — the very class the bacteria resist — or with polymyxins (35.6%), old and toxic drugs that have been revived precisely because little else works against these organisms [s1]. The reliance on carbapenems against carbapenem-resistant bugs is a marker of how few good options remain, not of clinical error.

Why this study is different

Most headline estimates of AMR mortality rest on modelling that combines microbiology data with broad assumptions, because prospectively tracking patients with confirmed resistant infections across many countries is expensive and slow. ACORN-HAI did the harder version: a multinational, multicentre cohort that consecutively enrolled patients and followed them to a hard 28-day endpoint, then separated crude from attributable mortality [s1]. The result is a firmer read on the true cost of these infections in the region where the resistant Gram-negative burden is heaviest.

The picture it produces is stark. In these Asian hospitals, resistance is the majority state of serious health-care-associated infection, not a minority complication; it concentrates in the Gram-negative pathogens with the emptiest drug cupboards; it kills a substantial share of those it infects even after accounting for how sick they already were; and it does so most in poorer countries and, disturbingly, among children and working-age adults [s1]. The authors' conclusion is that AMR bloodstream and ventilator infections impose a burden that the region's infection-control and stewardship systems are not yet containing — a burden this study, unusually, has measured directly rather than inferred.

This article is informational and is not medical advice.

Sources

• Severe health-care-associated infections and antimicrobial resistance in an Asian Surveillance Network (ACORN-HAI): a multicentre, prospective cohort study — The Lancet Infectious Diseases, 21 August 2026.