Prevalence and Associated Mortality of Infections by Multidrug-Resistant Organisms in Adults Intensive Care Units

Methods Study design PREVAR (PREValence of infection by MDR-microorganisms in ARgentina) was a national, cross-sectional, 24-hour multicentre study designed and conducted by SADI and SATI in Argentine ICUs.

Participants Sites were recruited to participate via announcements in national meetings, official societies' websites, social media, three online meetings explaining the project, and emails to all society members. Hospitals were registered on a secure website, where local investigators recorded main site characteristics in an electronic form.

Patient data were entered in an electronic case report form (CRF) using the Research Electronic Data Capture (REDCap) database (appendix 2). Local researchers were trained through online meetings on how to fill out the CRF. A centralized data center was responsible for managing the database and monitoring data quality in close contact with local researchers to minimize the occurrence of missing data. A dedicated email address was created to streamline communication.

Individual patient data were anonymised by assigning a numerical code to each case and recorded in order of admission.

Each local institutional review board approved the study and established the requirement for informed consent.

The PREVAR study protocol is available in appendix 3. All adult patients, 18 years and older, present on a participating ICU on the study day during the 24-h period (beginning at 8.00 a.m.) were included. The exact date could be selected from November 24-28, 2023, to ease logistics for individual sites. The patients present on day 1 in the ICU were available for inclusion in the cohort, which was then followed for mortality. There were no exclusion criteria.

Procedures The CRF included baseline and demographic characteristics of the study participants: date of hospital and ICU admission, age, gender, comorbid conditions, APACHE II and SOFA scores registered on admission; risk factors for MDRO infection during the previous six months, such as previous admission to a hospital, colonisation by Enterobacterales spp (only colonisation by CPE was measured, as it is the only type of colonisation usually monitored in Argentinian ICUs) and antibiotic utilization; type of admission (medical, elective or emergency surgery) was also collected. Data regarding colonisation by carbapenemase-producing Enterobacterales during the period occurring between hospital admission to enrollment was documented, whether colonisation had been detected in the hospital or in the ICU. SOFA score was also calculated on the day of enrollment.

Patient clinical status regarding infection was recorded by the treating physician according to the SEPSIS-3 definitions as: without infection, with infection but no sepsis, sepsis, and septic shock.10 In addition, infections were considered as definite (microbiologically confirmed), and probable or possible, according to the International Sepsis Forum definitions (ISF) for pneumonia, bloodstream infections (including infective endocarditis), intravascular catheter-related sepsis, intra-abdominal infections, urosepsis, and surgical wound infections. 11 In the cases where culture results were pending, categorisation was reviewed with the main investigators when the results were made available. Definitions of infection sites were those provided by the ISF. 11 Infections were registered as community-acquired, ICU-acquired, hospital non-ICU acquired, or as originating in long term care facilities.

According to microbiological findings, infections were recorded as MDRO or non-MDRO. MDRO's of interest were A.baumannii, difficult-to treat P. aeruginosa, carbapenemase-producing enterobacterales, extended spectrum β-lactamase producing microorganisms (ESBL), vancomycin-resistant enterococci, and methicillin-resistant S. aureus (MRSA) If available, further diagnostic methods were performed to further classified them as: KPC carbapenemase-producing enterobacterales (KPC), metallo-beta-lactamases (MBL), oxacillinases (OXA), extended spectrum β-lactamase (ESBL) producing organisms and AMPC beta-lactamases. The group of infections produced by non-MDRO for this study included infections by S.pneumoniae, S.pyogenes, S.aureus, E.Coli,, coagulase-negative Staphylococcus, Proteus sp., KES (Klebsiella-Enterobacter-Serratia-Citrobacter (KESC) group, C. difficile, and others.

The prescribed antimicrobial treatment was classified as empiric or targeted, according to the presence or not of positive cultures; for targeted treatment as adequate or inadequate, according to antimicrobial susceptibility testing; as utilizing novel or "old" antibiotics according to the European guidelines for the treatment of infections caused by multidrug-resistant Gram-negative bacilli 12; and as receiving Access, Watch and Reserve antibiotic groups according to the AWaRe WHO classification.) 13 Patients were followed until ICU discharge or death in the ICU.

Hospitals were classified as general or specialized, and as public or private. The number of total hospital beds were recorded as well as ICU beds, and the availability of certain relevant resources, such as infection control committee and antimicrobial stewardship program. Additionally, MDRO surveillance for infections and colonisation and how frequently it was performed (weekly, at ICU admission and/or before surgery) was detailed. Methods used to detect mechanisms of bacterial resistance were recorded as phenotypical, molecular or immunochromatographic.

The protocol of the study, the CRF and all the definitions appear in the Operations Manual (appendix 4).

Outcomes The main outcome measure was the prevalence of infection by MDRO. Secondary outcome measures were overall ICU mortality and its independent determinants, 28-day ICU mortality, determinants of infections by MDRO, prevalence of colonisation by CPE, and of infection in general, and ICU length of stay. Data was censored for 60 days.

Statistical analysis Prevalence was calculated as the number of infections caused by MDROs divided by the number of patients in the ICU on the day of the study.

Variables are reported as absolute numbers and percentages, or as medians and interquartile ranges (IQRs). Differences in recorded variables between patients with MDRO and non-MDRO infections, between survivors and non-survivors, and between patients with and without infection were analyzed using the χ² test or Fisher's exact test for categorical variables, and the t test or Wilcoxon rank-sum test for continuous variables, as appropriate.

To estimate the associations of patient characteristics, ICU organizational factors, and hospital characteristics with infection by MDRO and ICU mortality, a mixed-effects model was performed. These models have a hierarchical structure with patients (level 1) nested within hospitals (level 2). This nesting accounts for the potential correlation of outcomes within the same hospital. Random Effects: Hospitals were treated as random effects to account for the potential correlation of outcomes within the same hospital and to capture variability between hospitals through random intercepts Fixed Effects: Patient characteristics and ICU organizational factors were included as fixed effects, as these are the primary predictors of interest.

Only variables with a P value <.20 in the bivariable analysis were introduced into the final model. The results for fixed effects are reported as odds ratios (ORs) with 95% confidence intervals (CIs). Random effects, or measures of variation, are reported as the variance, its standard error (SE), and the median OR. The statistical significance of covariates was determined using the likelihood ratio test.

All reported p values are two-sided, with a p value <0.05 considered statistically significant.

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.