Contact precautions alone stopped ESBL spread in ICU study, no single rooms needed

Isolating patients who carry drug-resistant gut bacteria in single rooms has long been a cornerstone of intensive care infection control, but a prospective observational study from a Swedish university hospital suggests that careful contact precautions alone may be enough to stop patient-to-patient spread of extended-spectrum beta-lactamase-producing Enterobacterales, or ESBL-E, in an endemic setting with short stays. In the study, published in Annals of Intensive Care, researchers at the Karolinska University Hospital in Huddinge followed every admission to a ten-bed general intensive care unit for roughly 19 months, screened patients for ESBL-E on arrival, and then used whole-genome sequencing to check whether any bacteria found later in roommates matched the strains carried by their co-treated index patients. The verdict was unambiguous: not a single case of genomically confirmed cross-contamination occurred.

The question matters because international guidelines diverge sharply on how to manage carriers of multidrug-resistant organisms in intensive care. Swedish national guidelines and quality indicators, effective until 1 January 2025, mandated single-room isolation for all multidrug-resistant organism carriers to prevent cross-transmission. Guidelines from the United States Centers for Disease Control and Prevention recommend contact precautions for all patients colonized or infected with ESBL-producing organisms, along with single-room isolation when available. European ESCMID guidelines recommend the same procedures, but only in epidemic settings. The authors note that the CDC and ESCMID recommendations have mostly been studied during epidemic outbreaks rather than in endemic intensive care settings, which is precisely why the question of whether single rooms are truly necessary in routine practice has remained disputed.

The study was conducted in a ten-bed intensive care unit at Karolinska University Hospital, Huddinge, a tertiary and quaternary care hospital specializing in cell and organ transplantation and extensive upper abdominal surgery, including liver, pancreatic and esophageal procedures. The unit therefore handled a wide range of medical and surgical patients, including children. It comprised three units of four beds each, plus three isolation rooms, and patients were normally cared for in four-bed rooms with a nurse-patient ratio of 0.8 and an assistant nurse-patient ratio of 0.6. All patients enrolled in the study, including those colonized or infected with ESBL-E, were treated in these shared four-bed rooms under contact precautions.

Under the study protocol, contact precautions meant strict hand hygiene compliance, wearing gowns and gloves, and using disposable, single-use, non-critical care equipment. Environmental management followed local protocols, with daily cleaning of all surfaces surrounding the patient. On discharge, cleaning was performed with disposable cleaning utilities, and all single-use equipment, including pillows, was discarded. Routine cleaning and disinfection used detergent-based cleaning agents and disinfectants containing surfactants. These measures replaced the unit’s previous routine of placing known ESBL-E carriers in single rooms, which was the practice change the study was designed to evaluate.

Surveillance was intensive. Rectal swabs, along with swabs from drainages and intravenous catheters, were taken from all patients at admission to detect ESBL-E. Any patient who was a known ESBL-E carrier, regardless of a negative admission screen, or who tested positive at admission, was designated an index patient. Every patient co-treated in the same room as an index patient was then swabbed again at discharge. A new acquisition was defined as a negative screening result at admission followed by a positive screening at discharge. Whenever such an acquisition appeared, the bacterial isolate underwent whole-genome sequencing, and bioinformatic comparison with the index patient’s strain either verified or refuted cross-contamination. Cross-contamination was confirmed only when sequencing showed a close genetic relationship; low genetic relatedness ruled it out.

The study began on 4 November 2019 and ended on 30 June 2021, and the SARS-CoV-2 pandemic shaped its course. It was paused during Stockholm’s devastating first wave, from 17 March to 31 August 2020, and again briefly from 1 January to 22 January 2021 while the ethics permit was renewed. Of 1042 patients admitted during the study period, 129 were children under 20 years of age. Eighty-four patients, about 8 percent, were excluded because they required isolation in single rooms for other reasons, such as immunodeficiency, Clostridioides difficile infection, or carriage of carbapenemase-producing Enterobacterales, MRSA or VRE. Another 53 patients, about 5.5 percent, were excluded for missing admission swabs, leaving 905 patients in the analysis.

Among those 905 patients, 83, or 9 percent, were known ESBL-E carriers or tested positive at admission screening. Of the 822 patients who were ESBL-E-negative at admission, 356 were co-treated with an index patient during their intensive care stay, and 139 of those had exit samples taken according to protocol. Fifty-six exit samples were unavailable because the patients died in the unit, 38 were missed because of discharges within 24 hours, and the remainder were either improperly collected or lost. Only three patients from the co-treated group tested positive at discharge and were counted as new acquisitions, an acquisition incidence of 2 percent among co-treated patients. Two of these acquisitions involved Escherichia coli and one involved Salmonella enterica serovar Typhimurium. Crucially, genomic sequencing of the isolates from all three new acquisitions provided no evidence of cross-transmission from the index patients they had shared rooms with. The most common ESBL-producing species overall was E. coli, followed by Klebsiella pneumoniae.

The authors are candid about the study’s limitations. It was observational and conducted at a single center. A significant number of discharge swabs were missing or incorrectly collected, mostly toward the study’s end, a period that overlapped with the second and third pandemic waves in the Stockholm region. Although the unit had paused the study during the first wave, during the later waves it expanded from ten beds to a maximum of 30, dropping the nurse-patient ratio from 0.8 to 0.5. Continuous pressure on regular staff, combined with occasional inexperienced personnel, meant discharge routines were sometimes not maintained. The researchers also point to the unit’s short median length of stay as a limitation for generalizability: the median stay among index patients was 2.8 days, compared with 7 days reported by Alves and colleagues and 13 days reported by Barbier and colleagues, while the median length of stay in Swedish intensive care units during 2024 to 2025 was 2.51 days. Because a study by Alves and colleagues showed that patients who acquired ESBL had significantly longer intensive care stays, the short stays in this unit may partly explain the low acquisition rate.

Other contextual factors likely contributed as well. Antibiotic use among all patients admitted to the unit was 43.2 percent, but data on antibiotic treatment before admission were missing, even though research by Tacconelli and colleagues has associated antibiotic therapy during hospital stay with an increased risk of acquiring ESBL-E for up to seven days after treatment, with cephalosporin monotherapy showing the most significant effect. The authors also emphasize colonization pressure, the prevalence of resistant bacteria in the environment, which raises acquisition risk through environmental contamination and transient hand carriage among healthcare workers. Sweden’s low background prevalence of ESBL-E, with 4.7 percent carriage reported in a nationwide 2012 to 2013 study, means lower colonization pressure than in many other European countries. Adherence to hand hygiene was not collected during the study period, though it measured about 65 to 70 percent when assessed from 2022 to 2025, and environmental hygiene protocols were strictly followed. Counting all previously known carriers as index patients, even those with negative admission screens, may also have reduced acquisitions per index patient, since fecal carriage can persist for a year and negative samples do not necessarily indicate the end of carriage.

The findings align with earlier work showing minimal cross-transmission in both intensive care and non-intensive care settings. Repessé and colleagues found an ESBL-E acquisition incidence of 4.1 percent with only two cross-transmissions among 470 patients in a 12-bed French unit with no single rooms, where ESBL-E carriage prevalence was 13.2 percent. Tschudin-Sutter and colleagues identified just two cross-transmissions among 133 patients at a large Swiss tertiary care hospital, resulting in colonization without infection. Alves and colleagues confirmed only one cross-transmission among 19 acquisitions in 309 patients in a French medical intensive care unit, where admission carriage was 8 percent, closely matching the 9 percent seen in the Swedish study. Beyond resource economics, the authors note that single-room isolation is itself associated with an increased risk of adverse events and reduced contact with healthcare workers, so its benefits should outweigh its risks. They conclude that tailoring recommendations to the specific multidrug-resistant organism and to whether the unit faces an endemic or epidemic situation may offer a more constructive and cost-effective approach to reducing new ESBL-E acquisitions during intensive care, while cautioning that the acquisition of ESBL-E is multifactorial and that contact precautions are only one of several contributing factors.

Subject of Research: Whether contact precautions without single-room isolation prevent cross-contamination of ESBL-producing Enterobacterales in an intensive care unit

Article Title: Contact precautions prevent cross-contamination of extended-spectrum beta-lactamase-producing Enterobacterales in an intensive care unit: a prospective observational study

Article References: Milerad, N., Öhman, C. A., Fröding, I., Giske, C. G., & Castegren, M. (2025). Contact precautions prevent cross-contamination of extended-spectrum beta-lactamase-producing Enterobacterales in an intensive care unit: a prospective observational study. Annals of Intensive Care, 15(1), Article 193. https://doi.org/10.1186/s13613-025-01589-y

Image Credits: AI Generated

DOI: 10.1186/s13613-025-01589-y

Keywords: ESBL-producing Enterobacterales, contact precautions, intensive care unit, infection control, whole-genome sequencing, multidrug-resistant organisms, single-room isolation, colonization pressure, prospective observational study, antimicrobial resistance, Karolinska University Hospital, ICU length of stay