Carbapenem-resistant Enterobacteriaceae infection and associated factors in intensive care unit: a cross-sectional study

A cross-sectional study was conducted in the ICU of the tertiary-level hospital. The inclusion criteria comprised patients aged 18 and above who were admitted to the ICU from July 1, 2022, to December 31, 2022, and who underwent bacterial culture and identification testing. The patients are exclueded from the study if they lacked sufficient information about the tested results, such as the type of specimen infected with CRE, species CRE, and date of sample collection; the socio-demographics; or the treatment process in the ICU, as well as those non-compliant with treatment and arbitrarily ending the course of treatment without the consent of the doctor were excluded from the study.

The sample size was determined using the formula for estimating a population proportion. The study used the expected prevalence p=0.541 based on an Egyptian ICU study conducted from 2011 to 2017 on CRE epidemiology, using national healthcare-associated infection (HAI) surveillance data [13] and the margin of error d=0.05. Together with the expected 10% sample loss, the minimum total sample size to collect was 425 patients. The total number of patients admitted to the ICU from July 1, 2022, to December 31, 2022, was 455, which was approximated by the initial expected sample size of 425. So, the study used the method of collecting all data that met the inclusion and exclusion criteria for patients in the ICU during this time period.

The study utilized patients’ medical records to collect data, which involved completing an information collection form. This form included variables such as age, sex, source of patient, treatment department before admission to the ICU, hospitalization time within the last 3 months, comorbidities, CRE colonization, APACHE II score, sequential organ failure assessment (SOFA) score, Charlson comorbidity index (CCI) score, hospitalization time before CRE infection, surgery, HAI status, invasive procedure, exposure to antibiotics, treatment outcomes, hospitalization time, and CRE infection status.

Data were collected from July 2022 onward and included in the first ICU admission only. Each patient underwent daily monitoring from the ICU admission untill discharge or death at the hospital. Comprehensive patient information was collected from admission to the commencement of follow-up during this period.

As per CDC definition, CRE refers to an enteric bacteria that is resistant to at least one of the carbapenem antibiotics (minimum inhibitory concentration [MIC]≥4 μg/mL for Doripenem, Meropenem, or Imipenem, or ≥2 μg/mL for Ertapenem) or if they produce carbapenemase. Proteus spp., Morganella spp., and Providencia spp. may exhibited an elevated MIC for Imipenem. Therefore, to determine whether these microorganisms meet the criteria for CRE definition, the results of Ertapenem, Meropenem, and Doripenem should be considered [14].

Following the CDC’s HAI monitoring guidelines, a patient is classified as having a CRE infection based on the results of bacterial culture and identification testing as follows: for blood samples, the patient has a positive result (R: resistant) with ≥1 CRE confirmed CRE infection; for urine, sputum, and other fluid specimens, patients with positive results (R: resistant) with ≥1 CRE and a bacteria count>105 CFU/mL are identified as having a CRE infection.

The type of specimen infected with CRE is a substance taken from the patient’s body, such as sputum, blood, urine, and fluid, to perform tests and help diagnose diseases. The species CRE is a bacterium belonging to the group Enterobacteriaceae that resists Carbapenem recorded in the culture and identification of bacteria test results such as Escherichia coli, Klebsiella pneumoniae, Enterobacter spp., Serratia spp., and Proteus mirabilis. For each of thease variables, the study recorded only one occurrence of the specimen or bacteria. If a patient underwent culture at the same location more than once with consistent results, subsequent occurences were not recorded. This approahc ensured that results were not duplicated for the same patient.

This rate was calculated with the following formula:

The source of the CRE infection was classified in accordance with the CDC’s HAI monitoring criteria. This healthcare-associated CRE infection in the ICU was documented when a patient contracted a CRE infection during treatment at the ICU department and appeared 48 hours after the patient’s admission to the department. Conversly, a non-healthcare-related CRE infection in the ICU was identified when a patient developed a CRE infection either prior to ICU admission or more than 48 hours after transfer to another department or discharge from the hospital. This classification was based on the date of sample collection recorded in the test results.

Selection bias was mitigated through the compilation of a comprehensice list of patients aged 18 and above receiving treatment at the ICU from July to December 2022. Strict adherence to predefined research inclusion and exclusion criteria ensured clarity and simplicity. Information biase was controlled by clearly defining and specifying variables in the information collection form, which was desinged with precise objectives, clear language, ease of comprehension, and structured format. Investigators underwent thorough training and evaluation before official data collection commenced. following the data collection, 20% of the total collected forms were randomly selected for verification to completeness and accuracy by the researcher. These forms were also coded to facilitate management, data entry, data analysis, and report writing.

Data entry was performed using Epidata, and data analysis was carried out using STATA. Qualitative variables were decribed using frequency and percentage. Normally distributed quantitative variables were summarized using mean and SD, whereas non-normally distributed quantitative variables were descrived using median and interquartile range.

The Chi-square test and Poisson regression model were employed to assess the association between each factor and CRE infection. If more than 20% of the total cells had expected value less than 5, the Fisher test is used instead of the Chi-squared test. A significance level of p<0.05 was applied, and the prevalence ratio (PR) with a 95% confidence interval (CI) excluding 1 was considered statistically significant. Variables with p<0.05 in univariate analysis were selected for multivariable Poisson regression model to control for potential confounding factors and identify variables significantly associated with CRE infection.

The research report is adheres to the guidelines and checklist of strengthening the reporting of observational studies in epidemiology (STROBE), established by the UK EQUATOR Center, and updated on March 6, 2023 [15]. The checklist includes 22 components organized into six main sections: title and abstract, introduction, methods, results, discussion, and other information required for a cross-sectional research report and enhances the reliability of the findings presented.

The questions posed in this study were desinged to safeguard the patient’s automony and ensure their mental and physical well-being remained uncompromised. Personal information pertaining to the subjects were treated with confidentiality and was strictly utilzed for the purpose of the study alone. The analysis results derived from the study were employed solely for research-related objectives and were not utilized for any unrelated purposes.

The study protocol was approved in terms of research ethics by the Ethics Committee for Biomedical Research of the University of Medicine and Pharmacy at Ho Chi Minh City, No. 164/HĐĐĐ-ĐHYD, signed on February 14, 2023.

The study initially involved 455 patients admitted to the ICU from July to December 2022. However, four patients were excluded for not meeting the inclusion and exclusion criteria. This included one patient lacking information about their ICU treatment process and three patients who were still hospitalized at the time of sample collection. Consequently, the final anlaysis included 451 patients who met the criteria, meeting the minimum size requirement for the study. Therefore, the response rate for the initial expected sample size was 100% (Fig. 1).

Fig. 1. Flow diagram.

Download Original Figure

This study populaiton had a median age 67 years, with majority being male (56.5%). A significant proportion of patients originated from home or the community, and they had received treatment in the inpatient and emergency departments before ICU admisison. The median hospitalization duration within the last 3 months was 3 days. Most of the patients presented with comorbidities, with a median of 7 comorbidities (per patients?). Besides the main disease, the patients had cardiovascular disease, diabetes, and chronic kidney disease. The CRE colonization screening results were mostly negative (74.3%). The medians of APACHE II, SOFA, and CCI scores were 19, 5, and 5, respectively (Table 1).

The results showed that the median hospitalization duration prior to the CRE infection was 12 days. More than half of the patients had no surgery before determining the CRE infection. The prevalence of HAI was not high; however, patients with HAI mostly had bloodstream infections and ventilator-associated pneumonia. Most patients admitted to the ICU used invasive procedures such as endotracheal intubation, central venous catheters, arterial catheters, and urinary catheterization, which accounted for the largest proportion of invasive procedures. Most patients were exposed to antibiotics. The most commonly used group of antibiotics was ß-lactam, followed by Peptides, Quinolones, and at least Tetracyclines. The outcomes for patients were largely mitigated, with the median hospitalization time being 17 days (Table 2).

The results showed that the prevalence of CRE infection was 28.8%. The most common specimens and species of CRE were sputum and K. pneumoniae. More than half of patients with CRE infections were related to health care in the ICU (Table 3).

In the univariate anlaysis, several factors showed association with the rate of CRE infection. To mitigate potential confounding effects, the study conducted multivariable analysis using a multivariable Poisson regression model. Variables with a p-value<0.05 in univariate analysis were incorporated into the multivariable model. These variables include age, hospitalization duration within the last 3 months, number and types of comorbidities, neuropathy, screening for CRE colonization, APACHE II score, SOFA score, types of HAI, surgical site infection, types of invasive procedures, intubation, central venous catheter, arterial catheter, urinary catheterization, types of antibiotics: Aminoglycosides, Peptides, and Oxazolidinones; treatment outcomes; hospitalization time.

Results from the model showed that factors such as hospitalization time within the past 3 months, APACHE II score, SOFA score, surgical site infection, and hospitalization time were indeed associated with CRE infection.

Specifically, if the patient has 1 more day of hospital stay within the past 3 months, the prevalence of CRE infection will increase by 1% (p=0.026; PR=1.01; 95% CI 1.01–1.02). If patients’ APACHE II score increases by 1 point, the prevalence of CRE infection will increase by 3% (p=0.022; PR=1.03; 95% CI 1.01–1.06). If patients’ SOFA score increases by 1 point, the prevalence of CRE infection will decrease by 8% (p=0.042; PR=0.92; 95% CI 0.85–0.99). Patients with surgical site infection exhibited a 1.4 times higher prevalence of CRE infection compared to those without surgical site infection. The observed difference is statistically significant, with p-value=0.007 (PR=1.04, 95% CI 1.09–1.77). For each additional day of hospital stay, the prevalence of CRE infection is expected to increase by 1% (p<0.001; PR=1.01; 95% CI 1.01–1.02) (Table 4).