Carbapenem-Resistant Gram-Negative Bacteria Colonization in Immunocompromised and Critically Ill Pediatric Patients: Prevalence, Risk Factors, and Outcomes

Abstract

Background:

Carbapenem-resistant Gram-negative bacteria (CRGNB), including carbapenem-resistant Enterobacteriaceae (CRE) and Acinetobacter baumannii (CRAB) poses significant risks to immunocompromised and critically ill patients.

Methods:

A cross-sectional study was conducted at a tertiary care university hospital in Bangkok, Thailand (May 2023-October 2024). Hospitalized patients under 20 years of age who were immunocompromised or critically ill were enrolled. Rectal swabs were collected within 48 hours of admission.

Results:

Among 302 participants, 29 (9.6%) were colonized with CRGNB, including 19 (65.5%) with CRE and 10 (34.5%) with CRAB. Independent risk factors included recent surgery (adjusted OR 7.6, P = .006), hospital referral (aOR 3.03, P = .043), and younger age (aOR 0.86, P = .02). Colonized patients had longer hospital stays (median 12 days vs 6 days, P = .001) and higher mortality (17.2% vs 5.5%, P = .022).

Conclusion:

CRGNB colonization was identified in nearly 10% of high-risk pediatric patients and associated with unfavorable outcomes. Early identification and targeted infection control are essential.

Introduction

Antimicrobial resistance is a significant global public health threat with profound implications for health, mortality and economic stability. The 2024 update of the Bacterial Priority Pathogens List maintains Gram-negative bacterial pathogens in a critical status. Specifically, carbapenem-resistant Gram-negative bacteria (CRGNB), including carbapenem-resistant Enterobacteriaceae (CRE) and carbapenem-resistant Acinetobacter baumannii (CRAB) have received the highest evaluations, underscoring the urgent need for novel therapeutic strategies. ¹

CRGNB are notable healthcare-associated pathogens associated with significant morbidity and mortality due to their resistance to most antibiotics, including carbapenems. ² Recent data from the Thailand One Health Dashboard on Antimicrobial Resistance indicate a persistent increase in carbapenem resistance. Specifically, resistance in A. baumannii increased from 68% in 2018 to 76% in 2021. Carbapenem resistance in Escherichia coli increased from 2.4% in 2017 to 4.6% in 2021. Similarly, Klebsiella pneumoniae experienced a rise in carbapenem resistance from 10.1% to 16% over the same period. These trends highlight ongoing challenges in controlling CRGNB despite robust infection control measures. ³

The clinical impact of CRGNB infections is profound. A meta-analysis by Falagas et al ⁴ indicates that patients with CRE infections experience a mortality rate twice that of patients with carbapenem-susceptible Enterobacteriaceae, frequently due to delays in receiving appropriate antimicrobial treatment. Consequently, CRGNB infections pose an urgent challenge, necessitating enhanced prevention and optimized treatment strategies.

Colonization by CRGNB, defined as the presence of bacteria without active disease or symptoms, significantly increases the risk of hospital-acquired infections (HAIs), prolonged hospitalization, and higher mortality rates. ⁵ The prevalence of CRGNB colonization at admission varies widely among pediatric patients, from 4% to 50%, depending on demographics and geographical factors. Risk factors for colonization include prior hospitalizations, previous carbapenem use, and prolonged hospital stays.^{6 -11}

Immunocompromised and critically ill patients are particularly susceptible to colonization and subsequent infections due to a combination of compromised immune responses and disrupted physical barriers from invasive devices such as central venous catheters (CVC) and endotracheal tubes. The widespread use of broad-spectrum antibiotics in these populations contributes to dysbiosis, fostering the growth of resistant organisms. Prolonged ICU stays and underlying comorbid conditions further compromise their immune defenses, and frequent invasive procedures heighten their risk of exposure to multidrug-resistant organisms.^10,11

While routine screening and isolation for CRGNB colonization upon admission have been effective in reducing HAIs, ¹² standardized surveillance guidelines have yet to be established. Active surveillance, especially targeted at high-risk populations, is crucial for early detection of colonization and can potentially reduce the risk of subsequent infections. This study aimed to determine the prevalence of CRGNB colonization among hospitalized immunocompromised and critically ill pediatric patients, identify associated risk factors, and evaluate the clinical outcomes of those colonized.

Methods

Study Design and Participants

This cross-sectional study was conducted at a tertiary care university hospital in Bangkok, Thailand, which includes general pediatric units (56 beds), hemato-oncology unit (12 beds), semi-intensive care unit (intermediate ward; IW; 9 beds), and pediatric intensive care unit (PICU; 8 beds). The study included immunocompromised and critically ill pediatric patients, aged 1 month to 20 years, admitted to these units between May 2023 and October 2024.

Critically ill patients were defined as those requiring admission to the PICU or IW due to acute illness necessitating life-sustaining interventions, including mechanical ventilation, vasoactive or inotropic support, or close hemodynamic monitoring. Immunocompromised patients were defined by 1 or more of the following criteria: receipt of chemotherapy or high-dose corticosteroids (prednisolone ≥ 20 mg/day or ≥2 mg/kg/day for those under 10 kg) for ≥14 days, use of other immunosuppressive agents, status as hematopoietic stem cell transplant or solid organ transplant recipients, presence of decompensated liver cirrhosis, primary immunodeficiency, or HIV infection with CD4+ levels < 15% in patients under 13 years or <200 cells/mm³ in those 14 years or older. Exclusion criteria included patients with pre-existing CRGNB colonization or infection prior to enrollment, patients receiving postoperative care following elective surgery who did not meet the definitions of critically ill or immunocompromised, and patients who declined informed consent.

Ethical Approval and Informed Consent

The study protocol was approved by the Institutional Review Board of the study site (COA. MURA2023/385). Written informed consent was obtained from the legal guardians of all participants before enrollment.

Specimen Collection and Laboratory Methods

Patients who met the inclusion criteria and provided informed consent had rectal swab cultures performed within 48 hours of admission to screen for CRGNB colonization. The targeted CRGNB were CRE and CRAB. Rectal swabs were collected using sterile techniques and immediately placed into Stuart transport medium for transport to the microbiology laboratory. The specimens were inoculated onto selective agar media designed to isolate CRE and CRAB. Plates were incubated at 35°C to 37°C in ambient air for 18 to 24 hours. Suspected colonies were identified using standard microbiological techniques, including biochemical testing and, when necessary, VITEK2 automated systems. Antimicrobial susceptibility testing was performed according to the guidelines established by the Clinical and Laboratory Standards Institute. All results were recorded and reported in the hospital’s electronic medical record (EMR) system. In cases where rectal swab surveillance detected CRGNB colonization, the infection control nurse was notified through EMR system. Notifications triggered immediate implementation of contact precautions, including patient isolation and enhanced hygiene measures, to prevent transmission and reduce the risk of HAIs.

Data Collection

Data for enrolled patients were obtained through a comprehensive review of EMR, including demographic data (eg, age, gender), baseline characteristics, and comorbidities. In addition, data on potential risk factors for CRGNB colonization were systematically obtained. These factors included referral status from other hospitals, prior hospitalizations, antibiotic use within the last 3 months, recent major surgery (within 3 months prior to hospitalization), and the presence of medical devices such as urinary catheters, CVC, or mechanical ventilators. The clinical outcomes of interest included the length of hospital stays (LOS) and in-hospital mortality rates. LOS was recorded as the total number of days from admission to discharge, while mortality was defined as all-cause death occurring during hospitalization.

Statistical Analysis

Statistical analysis was performed using STATA statistical software (version 18.0, StataCorp, College Station, TX, USA). Continuous variables were described as means with standard deviations (SD) for normally distributed data or medians with interquartile ranges (IQR) for non-normally distributed data. Categorical variables were summarized as frequencies or percentages. The prevalence of CRGNB colonization was reported as a percentage. Comparisons between colonized and non-colonized patients were performed to identify potential risk factors. Pearson’s chi-square test or Fisher’s exact test was used for categorical variables, while the Student’s t-test or Mann-Whitney U test was used for continuous variables, based on data distribution. Univariate analysis identified variables associated with CRGNB colonization. Variables with P-values ≤ .2 were entered into a stepwise logistic regression model. During model selection, variables with P-values < .1 were retained, and those with P-values ≥ .2 were excluded. This approach aimed to minimize overfitting while maintaining clinical and biological relevance. Statistical significance was defined as P-value < .05.

Results

Patient Demographics

A total of 302 patients were enrolled in the study. Of these, 196 (64.9%) were classified as critically ill, and 106 (35.1%) were immunocompromised. Within the immunocompromised group, the most common condition was the use of immunosuppressive drugs, affecting 17.6% of these patients. This group also included cirrhosis patients (8.6%) and solid organ transplant recipients (5.3%). The median age of participants was 4.5 years (IQR 1.7 to 8.8), and 54% were male. Referral from other hospitals was documented in 57.3% of participants, while 23.8% had been hospitalized within the previous 3 months. Notably, 26.8% of the patients had an endotracheal tube in place, and 11.3% had CVC. Twelve patients (4.0%) had undergone surgery within the past 3 months, including 5 gastrointestinal procedures (eg, portoenterostomy, liver transplantation, exploratory laparotomy), 4 neurosurgical procedures (eg, tumor removal, ventriculoperitoneal shunt placements), and 3 cardiothoracic surgeries (eg, pulmonary artery banding, thoracotomy). A significant portion of the cohort (62.3%) had been exposed to at least 1 antibiotic in the 3 months prior to admission. The most commonly used antibiotic classes were non-carbapenem beta-lactams (54%), followed by carbapenems (15.6%), and macrolides (14.2%). The baseline characteristics of the study population are summarized in Table 1.

Table 1.

Baseline Characteristics of 302 Participants.

Baseline characteristics (n = 302)	N (%)
Age (years, IQR)	4.5 (1.7, 8.8)
Male gender	163 (54)
Host status
- Critically ill	196 (64.9)
- Immunocompromised	106 (35.1)
Immunocompromised status
- Receiving immunosuppressive drugs	53 (17.6)
- Decompensated cirrhosis	26 (8.6)
- Solid organ transplantation	16 (5.3)
- Hematopoietic stem cell transplantation	7 (2.3)
- Others	4 (1.3)
Patient’s source
- Referral from other hospitals	173 (57.3)
- Previous hospitalized within 3 months	72 (23.8)
Presence of medical devices
- Endotracheal tube	81 (26.8)
- Foley’s catheter	62 (20.5)
- Central venous catheter	34 (11.3)
History of surgery within 3 months	12 (4)
Type of antibiotics
- Any antibiotics	188 (62.3)
- Non-carbapenem beta-lactams	163 (54)
- Carbapenems	47 (15.6)
- Macrolides	43 (14.2)
- Aminoglycosides	18 (5.9)
- Fluoroquinolones	17 (5.6)
- Colistin	7 (2.3)
- Others	50 (16.6)

Abbreviation: IQR, interquartile range.

Prevalence of CRGNB Colonization

Among the 302 participants, 29 were found to have CRGNB colonization in the intestinal tract, resulting in an overall prevalence of 9.6%. Of these colonized cases, CRAB was the most frequently detected pathogen, presented in 10 cases (34.5%). The remaining 19 cases (65.5%) were attributed to CRE. Within the CRE group, K. pneumoniae was the most prevalent species with 9 cases (31%), followed by E. coli with 6 cases (20.7%) and Enterobacter cloacae with 4 cases (13.8%).

Risk Factors for CRGNB Colonization

The risk factors for CRGNB colonization were analyzed by comparing 29 participants who tested positive for CRGNB colonization with 273 participants who did not exhibit colonization, as summarized in Table 2. Significant differences in baseline characteristics were observed between the 2 groups. CRGNB colonized patients were generally younger, had a higher referral rate from other hospitals, were more likely to have CVC, and had undergone recent surgery within the previous 3 months. Additionally, antibiotic exposure was significantly associated with CRGNB colonization. Prior use of non-carbapenem beta-lactams, carbapenems, fluoroquinolones, colistin, and aminoglycosides, were more common among those colonized with CRGNB.

Table 2.

Comparative Analysis of Participants with and Without CRGNB Colonization.

Risk factors	CRGNB positive	CRGNB negative	P-value
	(N = 29)	(N = 273)
Age (years, IQR)	1.8 (0.42, 4.75)	4.5 (1.67, 8.75)	.033
Male gender	15 (51.7)	148 (54.2)	.798
Immunocompromised status
- Receiving immunosuppressive drugs	8 (27.6)	45 (16.5)	.135
- Decompensated cirrhosis	3 (10.3)	23 (8.4)	.726
- Solid organ transplantation	0	16 (5.9)	.38
- Hematopoietic stem cell transplantation	0	7 (2.6)	1
Patient’s source
- Non-referral & no previous hospitalized	0	57 (20.9)
- Referral	21 (72.4)	152 (55.7)
- Previous hospitalized within 3 months	8 (27.6)	64 (23.4)	.023
Presence of medical devices
- Endotracheal tube	8 (27.6)	73 (26.7)	.922
- Foley’s catheter	4 (13.8)	58 (21.3)	.345
- Central venous catheter	8 (27.6)	26 (9.5)	.009
History of surgery within 3 months	6 (20.7)	6 (2.2)	<.001
Antibiotic exposure within 3 months
- Non-carbapenem beta-lactams	22 (75.9)	141 (51.7)	.013
- Carbapenems	10 (34.5)	37 (13.6)	.006
- Macrolides	4 (13.8)	39 (14.3)	1
- Fluoroquinolones	6 (10.7)	11 (4)	.003
- Colistin	4 (13.8)	3 (1.1)	.002
- Aminoglycosides	5 (17.2)	13 (4.8)	.020
- Others	8 (37.6)	42 (15.4)	.113

Abbreviation: IQR, interquartile range.

Univariate analysis identified factors significantly associated with CRGNB colonization, including younger age, recent surgery, prior antibiotic exposure (specifically beta-lactams, carbapenems, fluoroquinolones, aminoglycosides, and colistin), referral from other hospitals, and the presence of CVC. Multivariate logistic regression analysis, after adjusting for confounding factors, identified that recent surgery within the past 3 months (adjusted odds ration [OR] 7.6, 95% confidence interval [CI] [1.79, 32.30], P = .006), referral from other hospitals (adjusted OR 3.03, 95% CI [1.03, 8.91], P = .043), and younger age (adjusted OR 0.86, 95% CI [0.75, 0.98], P = .02) remained significantly associated with CRGNB colonization. The details of factors associated CRGNB colonization are shown in Table 3.

Table 3.

Univariate and Multivariate Analysis of Risk Factors Associated with CRGNB Colonization.

Risk factors	Univariate analysis		Multivariate analysis
	Odds ratio (OR) (95% CI)	P-value	Adjusted OR (95% CI)	P-value
History of surgery within 3 months	11.61 (3.46, 38.89)	<.001	7.6 (1.79, 32.30)	.006
Referral from other hospital	2.1 (0.49, 4.88)	.089	3.03 (1.03, 8.91)	.043
Age	0.85 (0.76, 0.96)	.009	0.86 (0.75, 0.98)	.020
Central venous catheter	3.62 (1.46, 8.98)	.006	2.15 (0.58, 8.06)	.254
Carbapenems exposure within 3 months	3.36 (1.45, 7.78)	.005	1.03 (0.0.31, 3.46)	.963
Fluoroquinolone exposure within 3 months	6.21 (2.11, 18.33)	.001	4.15 (0.81, 21.20)	.087
Colistin exposure within 3 months	14.4 (3.05, 67.98)	.001	3.20 (0.42, 24.26)	.260
Aminoglycosides exposure within 3 months	4.17 (1.37, 12.68)	.012	1.50 (0.37, 8.34)	.644

Clinical Outcomes of CRGNB Colonized

Among the 29 patients with CRGNB colonization, 3 patients (10.3%) developed infections caused by the same pathogen identified during colonization. Two of these patients developed septicemia due to K. pneumoniae CRE and subsequently died, while 1 patient developed pneumonia caused by CRAB but survived. In addition to the 2 CRGNB-related deaths, 3 additional deaths occurred due to other causes, resulting in an overall mortality rate of 17.2% (5/29) among CRGNB colonized patients. Patients with CRGNB colonization had a significantly longer LOS compared to non-colonized patients (12 days [IQR 8-25] vs 6 days [IQR 3-12]; P = .001). Additionally, the mortality rate was significantly higher among CRGNB colonized patients (17.2%) compared to non-colonized patients (5.5%; P = .022).

Discussion

This study identified a prevalence of CRGNB colonization of 9.6% among critically ill and immunocompromised pediatric patients. Independent risk factors for colonization included recent surgery, referral from other hospitals, and younger age. Additionally, CRGNB colonization was significantly associated with prolonged hospital stays and higher mortality rates compared to non-colonized patients.

CRGNB pose a significant global health challenge, particularly in critically ill and immunocompromised individuals, due to their association with increased morbidity and mortality.^{10 -15} Intestinal colonization with CRGNB is a well-established risk factor for subsequent infection^5,16 and standard empirical antibiotic regimens, such as antipseudomonal beta-lactams, are often ineffective against CRGNB. ¹⁴ Consequently, infections caused by CRGNB are associated with increased disease severity, higher healthcare costs, prolonged hospitalizations, and increased mortality rates.^8,10 Active surveillance for CRGNB colonization is therefore essential, as it facilitates early detection, enable timely infection control measures, and reduced transmission and infection rates.

The prevalence of CRGNB colonization observed in this study is consistent with prior reports from Thailand^9,16,17 but appears lower than those reported in studies from Vietnam (35.8%) ⁸ and China (28.9%) ¹¹ However, these comparisons are descriptive and not based on formal statistical testing. Differences may reflect variations in healthcare settings, patient demographics, infection control policies, and antibiotic stewardship practices. Given the heterogeneity in study design, inclusion criteria, and local epidemiology, such comparisons should be interpreted with caution rather than as definitive statistical differences.

Recent surgery within 3 months was a significant risk factor for CRGNB colonization, aligning with previous findings.^18,19 Surgical procedures, particularly gastrointestinal surgeries, disrupt the mucosal barrier and gut microbiota, facilitating colonization by multidrug-resistant organisms. Additionally, surgical patients frequently receive broad-spectrum antibiotics, exerting selective pressure that promotes resistant strains. Moreover, invasive procedures, mechanical ventilation, and CVC use increase the risk of nosocomial pathogen acquisition, while prolonged hospital stays heighten exposure risk. This is particularly pronounced in tertiary care settings, where critically ill and immunocompromised patients foster an environment conducive to the circulation of multidrug-resistant organisms.

Transfer from other hospitals was identified as an independent risk factor for CRGNB colonization, likely due to prior healthcare exposure and hospital-based transmission from contaminated environments or colonized patients.^20,21 Critically ill and immunocompromised patients frequently receive broad-spectrum antibiotics, particularly carbapenems, further increasing colonization risk. for suspected infections. Variability in pathogen prevalence, infection control, and antibiotic stewardship across hospitals underscores the need for active surveillance in transferred patients, while antibiotic de-escalation may help reduce CRGNB emergence.

Younger age was also a significant risk factor for CRGNB colonization, with limited prior studies supporting this association.^{22 -24} Previous study showed that CRE colonization predominantly affected children under 2 years, particularly those born prematurely, likely due to increased contact with caregivers and healthcare workers, facilitating transmission. In addition, younger immunocompromised and critically ill patients often experience longer hospitalizations and higher antibiotics exposure, further predisposing them to CRGNB colonization.

CRGNB colonization was associated with prolonged hospitalization and increased mortality rates, consistent with previous studies.^8,10 Longer hospital stays not only increase healthcare costs but also increases the risk of nosocomial transmission, ²⁵ further underscoring the importance of early detection and prevention. The incidence of CRGNB colonization progressing to CRGNB infections was 10.3%, lower than that reported rates in China (27.3%) ¹¹ and Vietnam (21.5%). ⁸ This discrepancy may be influenced by differences in infection prevention strategies and host factors, such as immune status, underlying comorbidities, and prior antimicrobial exposure. Effective isolation measures, including room segregation, can further reduce transmission and infection rates.

This study has some limitations. First, the relatively small sample size and modest number of CRE-colonized cases may have limited the statistical power identify additional associations and some estimates may be less precise. Second, as a single-center study conducted in a tertiary care university hospital, the findings may not be generalizable to other settings with different patient populations and infection control measures. Furthermore, the exclusion of Pseudomonas aeruginosa screening, due to its low prevalence in our setting and removal from the WHO 2024 priority pathogen list, may have underestimated the overall burden of CRGNB colonization. Despites these limitations, this study provides valuable insights into the prevalence, risk factors, and clinical impact of CRGNB colonization in high-risk pediatric patients. Future multicenter studies with larger cohorts are warranted to validate these findings and support the development of targeted prevention strategies.

Conclusions

This study highlights the prevalence, risk factors, and clinical impact of CRGNB colonization in critically ill and immunocompromised pediatric patients. Recent surgery, hospital referral, and younger age were identified as independent risk factors, emphasizing the need for targeted infection prevention strategies. Targeted surveillance and optimized antimicrobial stewardship are essential to reducing CR-GNB colonization and improving outcomes. Future research should focus on multicenter studies to refine risk stratification and intervention strategies.