Associated Factors for Bacterial Colonization in Patients Admitted to the Intensive Care Unit of the Clinical Hospital of Infectious Diseases

Abstract

Introduction This study aimed to identify isolates from colonization and assess the risk factors for bacterial colonization and the risk of death in patients admitted to the intensive care unit (ICU) of the Constanţa County Infectious Diseases Hospital between September 2017 and September 2019. Methods This was a retrospective case-control study in a single center that included all patients admitted to the ICU in Constanța, Romania, who underwent bacteriological screening upon admission and 7 days after admission, between September 2017 and September 2019. In total, 253 patients were included in this study. The nasal exudate, pharyngeal exudate, and rectal swab samples were screened. Results In this study, 253 patients were screened bacteriologically, of which 53 had bacterial colonization and 200 did not. Among the bacterial strains, Klebsiella spp. (43.39%) was the most frequently isolated. The predominant resistance mechanism detected in the bacterial isolates was extended-spectrum β-lactamase (ESBL). Multivariate analysis identified a Carmeli score of 3 as an independent risk factor for acquiring bacterial colonization in the ICU. The mortality rate of patients with bacterial colonization was 11.32% and 6% for the patients without colonization (p > 0.05). Conclusions Our study revealed an increased prevalence of Enterobacterales colonization in the ICU. Risk factors for acquiring bacterial colonization differed depending on the type of bacterial colonization, such as ESBL, carbapenemases, methicillin-resistant Staphylococcus aureus (MRSA), and vancomycin-resistant enterococci (VRE). An independent risk factor for acquiring bacterial colonization was the Carmeli score of 3.

Introduction

Bacteriological screening is usually performed in the intensive care unit (ICU) and beyond and is an effective method for limiting the occurrence of healthcare-associated infections (HCAIs) [1]. Therefore, it is important to understand the circulation of bacterial isolates in hospitals. Based on these data, doctors decide on empiric antibiotic therapy for patients, which is often life-saving. It is important to understand the difference between bacterial colonization and infection when antibiotic treatment is initiated, especially if it is a reserved antibiotic. In a study conducted in seven wards of five university hospitals in Italy, it was observed that over 30 days, four (9%) of 42 patients newly colonized with antibiotic-resistant bacteria were infected by the same bacteria. One conclusion of this study was that early identification of colonization by antibiotic-resistant bacteria during antibiotic therapy could target a high-risk hospitalized population that may benefit from interventions to decrease the risk of subsequent healthcare-associated infections (HAIs) [2].

Patients are known to have an increased risk of colonization in the ICU owing to multiple invasive maneuvers [1]. The risks of bacterial colonization, including multidrug resistance (MDR) in hospitalized patients, increase if they have undergone previous prolonged hospitalizations, frequent hospitalizations, prolonged antibiotic therapy, or abusive antibiotic therapy. The risks pertaining to the medical act are increased by multiple invasive procedures, mechanical ventilation, use of contaminated devices, non-compliance with asepsis and antisepsis measures, non-compliance with medical unit circuits, and non-compliance with surveillance, prevention, and control of medical infections [3,4].

Patients who are colonized with multidrug resistant organisms (MDRO) have an increased risk of developing an infection with these microorganisms. Thus, it can be inferred that this risk may also extend to the patient’s home when the patient remains colonized by MDRO after discharge [3,4].

Another study supports the fact that patients who are at high risk of acquiring multidrug-resistant organisms are seriously ill, immunocompromised, or have been hospitalized for long periods of time. Additional risk factors include prolonged use of antibiotics and contact with colonized patients or colonized/contaminated hands [5]; therefore, it is important to know the local circulation of germs and the risk factors for acquiring MDRO.

According to a report by the European Centre for Disease Prevention and Control (ECDC) in Romania in 2018, the levels of antimicrobial resistance (AMR) in Romania represent a serious concern. Reported levels of AMR are very high and/or increasing compared to most European Union/European Economic Area (EU/EEA) countries. The report also highlighted limitations in infrastructure in hospitals and a lack of training for healthcare personnel. Furthermore, Romania lacks robust data on AMR rates, HAIs, and antibiotic consumption on local level [6]. This study aimed to identify the strains of methicillin-resistant Staphylococcus aureus (MRSA), extended-spectrum β-lactamase (ESBL) producers, carbapenemase producers, and vancomycin-resistant enterococci (VRE) in our hospital and to identify the risk factors associated with acquiring strains with resistance mechanisms.

Methods

Study population

The approval of the data processing of the patients from the ICU of the Constanța Clinical Hospital for Infectious Diseases, Romania, from September 2017 to September 2019, was obtained. Patients hospitalized in the ICU who were bacteriologically screened upon admission to the ICU and 7 days after admission to the ICU according to the hospital protocol were divided into the case group (bacterially colonized patients) and the control group (non-colonized patients).

The ICU has 10 beds. This study included 253 patients admitted to the ICU who underwent bacteriological screening upon admission. Bacteriological screening was also performed seven days after admission in patients who remained in the ICU for this period.

The following inclusion criteria from the patients’ data were taken into consideration: hospitalization in the ICU for more than 24 hours, between 2017-2019 regardless of the pathology of admission, treatment administered, and epidemiological, clinical, and paraclinical data.

Exclusion criteria included patients admitted to the ICU who had not undergone bacteriological screening, those who were admitted to the ICU for less than 24 hours, those who were hospitalized once or several times in the ICU during hospitalization from the Infectious Disease Hospital, and those who were identified with the same MDR colonization, the same bacterial strain, and the same resistance mechanism.

Data collection

Data were collected from the hospital’s computer system. Demographic, clinical, and microbiological data were also obtained. The Carmeli score was used to detect the risk of colonization and bacterial infection by MDRO.

The criteria that composed the Carmeli score were obtained as follows: score 1 (being classified as community-acquired), score 2 (being considered as HAI), and score 3 (classified as nosocomial) [7].

The Charlson Comorbidity Index (CCI) is used to estimate the risk of death due to a comorbid disease and has been widely used as a predictor of prognosis and long-term survival. Thus, CCI is derived by summing the weights assigned to all comorbid conditions presented by patients, with higher scores indicating worse prognosis [8]. This score was also used to determine the risk of colonization in patients with high scores.

Other patient data such as comorbidities, previous hospitalization in the last 6 months or antibiotic treatment 6 months prior to admission to the ICU, current antibiotics including restrictive antimicrobials (colistin, linezolid, vancomycin, tigecycline, and carbapenems), antiviral and corticosteroid treatment, and the patients’ origin environment (urban or rural) were also analyzed.

This study was approved by the Ethics Committee of the Constanţa Clinical Hospital for Infectious Diseases (NR 1/07/01.2022), CODE F.05 PO.17.00-ACFOCG.

Microbiological data

The microbiological laboratory tests used to perform bacteriological screening in the ICU included phenotypic methods for bacterial identification. Chromogenic agar medium, the double-disc synergy test, and the Modified Hodge test were used.

From the rectal swab, Gram-negative bacilli (GNB) producing ESBLs were identified using selective culture media ESBL chrome agar (CHROMID ESBL, bioMérieux, France). ESBL production was confirmed using the double-disc synergy test. Escherichia coli ATCC 35218 was used in this study. For the quality control of Klebsiella spp., ATCC 700603 and ATCC BAA-2814 were used.

Detection of carbapenem-resistant Enterobacteriaceae (CRE) and carbapenem-resistant Acinetobacter baumannii (CRAB) isolated from rectal swabs was performed using carbapenemase-producing Enterobacteriaceae (CHROMID CARBA, bioMérieux) chromogenic agar medium and confirmed using the Modified Hodge test (MHT) or Rosco disc. Screening for carbapenemase production has a cut-off for an MIC of > 0.125 and screening with 10 µg disks of <28.

vancomycin-resistant E. faecalis and E. faecium (VRE) strains were detected in rectal swabs using a selective chromogenic medium (CHROMID VRE, bioMérieux). Screening for vancomycin resistance in enterococci has an MIC of 4 mg/L, and 10 µg discs of 12.

For patients who, during the screening of rectal swabs for CPE/VRE, had on chromogenic media the growth of some colonies characteristic of resistant strains that were further identified with VITEK 2/MALDI TOF, confirmation of resistance to antibiotics by performing ATB either by difusimetric or by microdilution (VITEK, bioMérieux) an attempt was made to identify a resistance mechanism by Rosco test, Hodge test.

Chromium MRSA or CHROM agar chromogenic media (CHROMID D MRSA, bioMérieux) were used to detect MRSA in nasal or pharyngeal exudates. Isolates were confirmed by diffusion discs using a cefoxitin or oxacillin disc. A quality disk (Staphylococcus aureus ATCC 29213) was used for quality control.

The bacterial isolates were identified using VITEK 2 (bioMérieux), (MALDI-TOF (MALDI TOF Autof MS 1000, Bruker, Germany), or disk diffusion. The antibiogram was interpreted according to the European Committee for Antimicrobial Sensitivity Testing guidelines (EUCAST) [9]. Antibiograms were not obtained for all bacterial isolates detected by screening.

A limitation of this study is that no molecular tests, chromogenic lysis tests, or fast immunochromatographic tests were used to confirm carbapenemase production because of the high costs.

Statistical analyses were performed using Microsoft Excel 2019 (Microsoft Corporation, USA) and SPSS Statistics (version 20.0, IBM Corp., USA). We used the Chi-square test as a categorical variable to compare the differences between the groups. Univariate analysis identified the factors associated with the risk of bacterial colonization. The relative risk (RR) for these patients was also determined. To identify independent risk factors, a multivariate analysis with binary logistic regression (LR) was performed.

Time-dependent covariate analysis was performed for the length of stay in the ICU of patients with colonization to determine if it was associated with an increased risk of death.

Statistical significance was set at p≤ 0.05. All the reported p-values were two-tailed.

Results

Prevalence and location of bacterial isolates

In the ICU ward of the Infectious Disease Hospital, 296 patients were admitted at least once during the same hospitalization period. Fifty-three (20.9%) of the 253 tested patients were found to be colonized with bacteria in bacteriological screening upon admission to intensive care.

A single bacterial isolate was detected in 37 (68.8%) patients. Most of these patients were identified with Escherichia coli, followed by Klebsiella spp., Staphylococcus aureus and Enterococcus faecium strains (

Table 1. Bacterial strains detected in bacteriological screening.

Bacterial strains
All microorganisms	N=53, N (%)
Klebsiella spp.	12 (22.6)
Escherichia coli	17 (32.1)
Staphylococcus aureus	6 (11.3)
Enterococcus faecium	2 (3.8)
Klebsiella spp. and Acinetobacter baumannii	1 (1.9)
Klebsiella spp. and Enterobacter cloacae	2 (1.9)
Klebsiella spp. and Staphylococcus aureus	1 (1.9)
Klebsiella spp. and Enterococcus faecium	3 (5.7)
Klebsiella spp. and Enterococcus faecalis	1 (1.9)
Escherichia coli and Staphylococcus aureus	1 (1.9)
Escherichia coli and Enterococcus faecium	1 (1.9)
Staphylococcus aureus and other Enterobacteriaceae	1 (1.9)
Staphylococcus aureus and Acinetobacter baumannii	1 (1.9)
Staphylococcus aureus and Enterococcus faecium	1 (1.9)
Klebsiella spp., Escherichia coli and Enterococcus faecium	1 (1.9)
Klebsiella spp., Enterococcus faecium and Enterococcus faecalis	2 (3.8)

). Thirteen (24.5%) patients had two distinct bacterial isolates, of which the most predominant associations were Klebsiella spp. and Enterococcus faecium. However, the bacteriological prevalence of patients who had three distinct bacterial isolates was detected in only three patients (5.7%) – Table 1. The most frequently detected bacterial strain was Enterobacterales (Table 1).

Regarding the location of bacterial isolates, statistical data showed that 42 (79.2%) of the bacterial isolates were detected in rectal swab samples, six (11.3%) strains were isolated from nasal exudate samples, five (9.4%) were isolated from both rectal swab and nasal exudate samples, and none were isolated from pharyngeal exudate.

The resistance mechanisms of bacterial isolates

Table 3 shows the resistance mechanisms detected, and it can be seen that the predominant mechanism of resistance during bacterial colonization is the production of extended-spectrum β-lactamases (ESBL), with 40 (75.5%) isolates detected. The second mechanism of resistance detected was the production of carbapenemases in 17 (32.1%) patients, and the third resistance mechanism detected was MRSA or VRE, with 11 bacterial strains detected (Table 2).

Table 2. Mechanisms of resistance in bacterial colonization detected in bacteriological screening.

Mechanism of resistance seen in bacterial colonization	(N=53), N (%)
ESBL	20 (37.8)
Carbapenemase	4 (7.6)
MRSA	6 (11.3)
VRE	2 (3.8)
ESBL/carbapenemase	7 (13.2)
MRSA/VRE	1 (1.9)
ESBL/carbapenemase/VRE	6 (11.3)
ESBL/VRE	3 (5.7)
ESBL/MRSA	4 (7.5)

ESBL – extended spectrum beta-lactamase; MRSA – methicillin-resistant Staphylococcus aureus; VRE – vancomycin resistant enterococci.

Table 2. Mechanisms of resistance in bacterial colonization detected in bacteriological screening.

Mechanism of resistance seen in bacterial colonization	(N=53), N (%)
ESBL	20 (37.8)
Carbapenemase	4 (7.6)
MRSA	6 (11.3)
VRE	2 (3.8)
ESBL/carbapenemase	7 (13.2)
MRSA/VRE	1 (1.9)
ESBL/carbapenemase/VRE	6 (11.3)
ESBL/VRE	3 (5.7)
ESBL/MRSA	4 (7.5)

ESBL – extended spectrum beta-lactamase; MRSA – methicillin-resistant Staphylococcus aureus; VRE – vancomycin resistant enterococci.

Risk factors associated with bacterial colonization

Based on the results of the univariate analysis, the risk factors associated with bacterial colonization are described in Table 3, Table 4 and Table 5.

Univariate and multivariate analyses revealed that a Carmeli score of 3 was a significant risk factor for acquired bacterial colonization. Patients with a Carmeli score of 3 points were twice as likely to acquire bacterial colonization (OR, 2.21; 95%CI, 1.10-4.46, p≤0.05). In the univariate analysis, patients who had bee exposed to hospitalization or antibiotics in the past 6 months were twice as likely to acquire bacterial colonization (OR, 2.52; 95%CI, 1.22-5.16, p≤0.05) – Table 3. Univariate analysis was performed using the Chi-square test, and multivariate analysis was performed using binary logistic regression.

Table 3. Characteristics of patients with colonization.

Table 3. Characteristics of patients with colonization.

Table 4 shows that there were 43 (81.1%) patients who had chronic diseases and 151 (75.5%) patients who did not have chronic diseases. Regarding the univariate analysis and the multivariate analysis, we notice that chronic diseases were not risk factors for the acquisition of bacterial colonization (Table 4).

Table 4. Chronic diseases.

Table 4. Chronic diseases.

Regarding the treatment of patients during hospitalization, a positive association was noted only between metronidazole treatment and bacterial colonization. Multivariate analysis of the treatments administered to patients during hospitalization revealed no independent risk factors for acquired bacterial colonization (Table 5).

Table 5. Treatment of hospitalized patients.

Table 5. Treatment of hospitalized patients.

Discussion

It is important to identify and detect microbial colonization with isolates such as VRE, MRSA, carbapenemase-producing, and ESBL-producing bacteria in patients admitted to the ICU. According to EARS NET 2019, antimicrobial resistance has increased. There are large variations depending on the geographical region, bacterial species, and antimicrobial treatment administered to the patients [10].

The species most frequently reported by the ECDC in 2019 were Escherichia coli (44.2%), Staphylococcus aureus (20.6%), Klebsiella pneumoniae (11.3%), Enterococcus faecalis (6.8%), Pseudomonas aeruginosa (5.6%), Streptococcus pneumoniae (5.3%), Enterococcus faecium (4.5%), and Acinetobacter species (1.7%) [10]. In our study, the report was not similar, Klebsiella spp. being the most frequent bacteria detected, followed by Escherichia coli, Staphylococcus aureus, Enterococcus faecium, Enterococcus faecalis, Enterobacter cloacae, Acinetobacter baumannii and other Enterobacterales. According to data published in Romania, our country is among the member states with the highest AMR levels in Europe [6].

According to international data, E. coli and Klebsiella spp. are the most common ESBL-producing bacteria. Studies have revealed that the prevalence of ESBL-producing Enterobacteriaceae has dramatically increased worldwide and has started to represent a global threat. The causes of this rapid increase are the spread of resistance mechanisms, excessive use of antimicrobial agents, and inadequate and ineffective control measures for the detected infections [11]. The findings of our study showed that ESBL is the most frequent mechanism of resistance in the ICU. Although our study did not pursue all predictable risk factors for acquiring bacterial colonization, it was shown that ESBL strains detected alone or in combination with other resistance mechanisms had a Carmeli score of 3 as an independent risk factor.

Carbapenem-resistant Enterobacterales (CRE) strains are known to be associated with a high mortality rate of up to 50% in hospitalized patients according to the Centers for Disease Control and Prevention (CDC), which requires a clinical laboratory to accurately identify CRE, alert medical personnel, and perform tests for the identification of carbapenemases [12,13]. In our study, colonizing strains producing carbapenemases were detected, although these were found in only a small percentage (32%) of the total detected bacterial strains. In a study conducted on bacterial strains from Romania, 65 of 75 carbapenem-non-susceptible isolates were found to be carbapenemase producers, highlighting the need to implement hospital infection control measures [14].

According to some reports, VRE causes multiple infections in hospitalized patients, and some of them die. Therefore, it is important to identify the risk factors for VRE in hospitalized patients. The CDC estimates that the risk factors for VRE are prior treatment with antibiotics, including vancomycin, for long periods of time, previous hospitalization, invasive procedures, and hospitalization in the ICU or in transplantation or oncology units [15]. In our study, the number of patients who were colonized with VRE was 22.6% of the total strains detected and 4.74% of the total bacteriologically screened patients, a percentage that was also low compared with the European level. However, it should be noted that the data in our study refer only to bacterial colonization without including VRE infections. In EU/EEA, the weighted average percentage of resistance to vancomycin in E. faecium was 18.3% in 2019, and in E. faecalis it remained low in most countries [10]; however, in Romania, data from 2019 showed an increased percentage, between 25-50% [10].

There is not enough data reported in Romania regarding antimicrobial resistance, HAIs, and antimicrobial treatment [8]. Bacteriological screening is not performed in all ICUs; thus, no bacterial isolates are known to cause severe infections associated with this medical act. In the study conducted by Pirii et al., it was noted that data were reported on burn patients who were colonized with carbapenemases as infectious consequences of the Colectiv fire disaster [16], a fact that also shows that in Romania antimicrobial resistance is increased, and their detection is not always feasible, possibly due to the fact that bacteriological screening is not performed in all patients who are at risk of MDRO.

In another article by Ungureanu et al., which was performed on patients from a regional hospital in Craiova, Romania, the number of patients with nasal and pharyngeal colonization detected in pharyngeal exudates was 400 (32.30%) and 48 (33.80%) of those with positive nasal exudates. The species that were isolated included Staphylococcus aureus in pharyngeal swabs in 67.21% of the cases and Staphylococcus aureus in nasal swabs in 75.41% of the cases, whereas other species that were detected included Klebsiella spp., Escherichia coli, Proteus spp., Enterobacter spp. and Pseudomonas spp. [17], which were not similar to those in our study.

In a study carried out in another county under similar conditions to our study, in a single specialized hospital and only in the ICU with a limited number of beds, a totally different prevalence was observed: 5% and 4% strains with MRSA, 24.1% with ESBL Enterobacterales, 10.7% with carbapenemases, and 13.4% strains with Enterococcus faecium resistant to vancomycin [18]. In our study, 40 (15.81%) isolates had ESBL, 17 (6.71%) carbapenemases, 11 (4.34%) MRSA, and 12 (4.74%) VRE.

In other international studies, it was observed that the risk factors for acquiring colonization with carbapenemase-producing Enterobacterales were female sex, presence of solid tumors, hematopoietic malignancy, immunodeficiency, diseases of the urinary system combined with the use of antibiotics, and bronchoscopy procedures [19,20,21] which are also different from those identified in our study.

Another study by Wyres et al. suggested that a more detailed assessment of risk species, especially the risk of transmission and infection rate, is needed to effectively guide rectal screening programs and limit MDR bacteria [22].

In 2017, a special ECDC report in Romania was published to discuss problems related to antimicrobial resistance, concluding that certain improvements were needed in our country to limit the use of antibiotics and prevent an increase in antibiotic resistance. Public awareness campaigns are recommended to limit the use of antibiotics and interdisciplinary collaboration between clinicians, epidemiologists, microbiologists, and veterinarians [6].

Thus, a robust and effective antimicrobial stewardship program is essential to minimize the risk of infection and implicit infection [23,24].

The limitations of the present study are related to the lack of knowledge of all the possible risk factors that can be encountered in an intensive care unit, including the lack of knowledge of compliance with hygiene measures, but also of other factors such as knowledge of the patient lounge and the prophylactic measures taken for each MDRO patient, such as failure to observe universal precautions.

The motivation for conducting this study is justified by the need to corroborate bacteriological screening data with that of bacterial infections with resistance mechanisms to initiate an antimicrobial stewardship plan. Thus, knowing the circulation of germs and risk factors would help clinicians stop the increase in infections associated with medical acts.

Conclusions

Our study revealed that Enterobacterales isolates represented the predominant bacterial colonization in the ICU of the Constanța Clinical Hospital for Infectious Diseases between September 2017 and September 2019. The independent factor that increased the risk of acquired bacterial colonization was a Carmeli score of 3 points. The mortality rate was not higher in colonized patients and was not dependent on the length of stay in the ICU. Thus, with the help of this study, a local guide with recommendations can be made for patients at an increased risk of MDRO.