INTRODUCTION
The recent increase in multidrug-resistant bacteria (MDR) in healthcare settings is recognized as a global public health problem. Antimicrobial resistance, which has become a natural phenomenon over time, could be accelerated by the misuse of antibiotics in animals and humans, leading to limited treatment options, causing life-threatening infections and increasing hospital costs1. In order to implement a multimodal approach, which guarantees the best treatment possible and prevents MDR bacterial infections, in 2015, the World Health Assembly adopted the global action plan on antimicrobial resistance2.
MDR bacteria are present in children and neonate hospitals, especially in intensive care units. A recent systematic review by Le Doare et al, which has been reporting MDR since 2001, described regional databases of Gram-negative bacteremia in children in low- and low-middle-income countries. In the review, 71,326 children from 30 studies were included, and Gram-negative bacteria were isolated from 4,710 (66.8%) out of 7,056 blood cultures. The review revealed that in neonates in Asian countries, the resistance of Klebsiella pneumoniae to cephalosporins and ampicillin was 84% and 94%, respectively, and in African countries, it was 50% and 100%, respectively. The authors observed that multidrug resistance, including resistance to ampicillin, chloramphenicol, and cotrimoxazole, was 30% (interquartile range [IQR], 0-59.6) in Asia and 75% (IQR, 30-85.4) in Africa3. Another study performed at a single neonatal intensive care unit (NICU) in Taiwan revealed that MDR Gram-negative bacilli (GNB) accounted for 18.6% of all neonatal GNB bacteremia, and drug resistance was more frequently reported in neonates with previous broad-spectrum antibiotic therapy4.
MDR, which causes healthcare-associated infections (HAI), is also a problem in Brazilian health units. Folgori et al, described HAI epidemiology in neonatal/pediatric intensive care units (NICU/PICU), as well as the impact of MDR in three pediatric hospitals being 1 from Italy and 2 from Brazil. The cumulative incidence of HAI in the different hospitals was 3.6/100 ICU admissions, and the crude 30-day fatality rate was 5.7/1,000 admissions. Of the 206 HAI episodes reported in Brazilian ICUs, 122 (60.1%) were due to MDR bacteria5.
Nosocomial infection and multidrug-resistant bacteria surveillance systems could be useful in the understanding of MDR epidemiology in children and neonates within a geographical region, as well as from a global perspective. It could also help in the recognition of prevalent pathogens based on the site of infection, in the optimization of resources, and in the stimulation of rational antibiotic use, while contributing to multimodal strategies aimed at reducing MDR rates6. Well-structured nosocomial infection surveillance systems, which are governmental and non-governmental initiatives, are described around the world7–11.
The aim of this study was to describe colonization and infection-causing MDR bacteria patterns in neonatal (NICU) and pediatric intensive care unit (PICU) patients, reported to the Rio de Janeiro State surveillance system.
METHODS
A cross-sectional survey was performed between 2014 and 2017 using electronic data from healthcare institutions with neonatal and/or pediatric beds, collected by the Rio de Janeiro State MDR bacteria surveillance system.
Surveillance system
The Rio de Janeiro State MDR surveillance system, which excludes multi-resistant fungi, is a web-based surveillance system created in 2013. Since then, it has become mandatory for hospitals located in the state to report any identified case of multidrug-resistant (MDR) bacteria. Data input in this electronic system required information on the identified MDR type (Gram-positive/Gram-negative), a description of its association with colonization or infection, and the site of infection (if present).
Inclusion/exclusion criteria:
All healthcare institutions that notified at least one case during the study period were included.
Definition of Multi-drug-resistant bacteria
The following MDR agents were reported: ESBL producing Klebsiella and E.coli, 3rd or 4th generation cephalosporins-resistant Enterobacteriaceae (Enterobacter/Proteus/Morganella and Citrobacter), carbapenem- or polymyxin-resistant Acinetobacter/Pseudomonas aeruginosa, carbapenem- or polymyxin-resistant Enterobacteriaceae (E.coli, Klebsiella spp, Enterobacter spp, Citrobacter spp, Serratia spp, Providencia spp, Proteus spp, Morganella spp, and Citrobacter), vancomycin-resistant Enterococcus, methicillin-resistant S. aureus, vancomycin-intermediate S. aureus (MIC 4-8 µg/ml), and vancomycin-resistant coagulase-negative Staphylococcus (MIC > 4).
Definition of Healthcare-associated infections
All infections were identified in accordance with the Brazilian National Health Surveillance Agency criteria (ANVISA) for neonates (0-28 days old) and children (>28 days-18 years old)12–13. Only infections presented and identified more than 48 h after admission to healthcare institution were considered HAI, otherwise, they were classified as colonization.
Variables of interest
The following variables were analyzed: age, total colonizations and infections reported per year, MDR bacterial class identified in colonizations and infections, and MDR bacterial class based on site of infection.
Data analysis
Patients were categorized as neonates or children, following their ages. In addition, MDR bacteria were categorized as Gram-positive or Gram-negative, and their colonization and infection frequencies were presented as percentages. Further, the frequencies of specific pathogens within the Gram-positive and Gram-negative groups were presented as percentages. Descriptive analysis was conducted using the Excel spreadsheet® version 2016. (Microsoft Corp., Redmond, WA, USA).
RESULTS
Between 2014 and 2017, 139 healthcare institutions (78 NICUs and 61 PICUs), reported 10,210 MDR bacteria cases, which included 9,261 colonizations (5,379, NICUs; 3,882, PICUs) and 949 infections (405, NICUs and 544, PICUs), as shown in Figure 1.

FIGURE 1: Multi-drug-resistant (MDR) bacteria colonizations in NICU and PICU patients, reported to the Rio de Janeiro State surveillance system (2014-2017).
The mean of colonization-causing MDR bacteria was 1,344.75/yr (1,075-1,788 range) in NICUs and 970.5/yr (656-1625 range) in PICUs. Compared with the different annual mean notifications, the number of colonizations in NICUs and PICUs in 2017 increased by 33% and 67.4%, respectively.
Table 1 shows the absolute numbers and frequencies of the main NICU and PICU colonization-causing MDR bacteria reported. ESBL producing Klebsiella sp and E.coli, and MRSA, were the most reported colonization-causing MDR agents in neonates and children; however, regarding relative frequency, CRGNB represented an important group in PICUs.
TABLE 1: Absolute number and frequency (%) of the main multidrug-resistant (MDR) bacteria colonizing NICU and PICU patients, reported to the Rio de Janeiro state surveillance system (2014-2017).
Bacteria | NICUs | PICUs |
---|---|---|
N = 5,379 | N = 3,882 | |
(%) | (%) | |
ESBL-producing Klebsiella sp/E.coli | 1983 | 1494 |
(36.9) | (38.5) | |
MRSA# | 1,197 | 1,075 |
(22.2) | (27.7) | |
Enterobacteriaceae* resistant to 3 rd or 4 th generation cephalosporins | 559 | 136 |
(10.4) | (3.5) | |
CRGNB** | 533 | 396 |
(9.9) | (10.2) |
#Methicillin-resistant S. Aureus, *(Enterobacter/Proteus/Morganella, Citrobacter), **CRGNB (Carbapenem-resistant Gram-negative bacteria, including Pseudomonas aeruginosa, Acinetobacter spp, and Enterobacteriaceae).
The mean of infection causing MDR bacteria was 101.25/yr (65-125) in NICUs and 136/yr (101-185) in PICUs, and when they were compared with the different annual means, the results revealed that the number of reported infections in NICUs and PICUs in 2017 increased by 23.5% and 36%, respectively, as shown in Figure 2. ESBL producing Klebsiella sp and E.coli were the most reported CLABSI- and surgical site infection (SSI)-causing MDR bacteria in NICUs, carbapenem-resistant Gram-negative (CRGNB) bacteria were the most reported catheter-associated urinary tract infection (CAUTI)- and ventilator-associated pneumonia (VAP)-causing MDR bacteria, and MRSA was the most reported neonatal sepsis-causing MDR bacteria (Table 2).

FIGURE 2: Infections by multi-drug-resistant (MDR) bacteria in NICU and PICU patients, reported to the Rio de Janeiro State surveillance system (2014-2017).
TABLE 2: Multi-drug-resistant (MDR) bacteria in neonates admitted to NICUs, according to infection site, reported to the Rio de Janeiro State surveillance system (2014-2017)
CLABSI | VAP | CAUTI | SSI | Neonatal sepsis | ||
---|---|---|---|---|---|---|
(%) | (%) | (%) | (%) | (%) | ||
Gram-positive | MRSA# | 25 | 13 | – | 3 | 14 |
(16) | (20) | (37.5) | (43.8) | |||
VISA## | 3 | – | – | – | – | |
(1.9) | ||||||
VRE### | 1 | – | 1 | – | – | |
(0.6) | (2.8) | |||||
Gram-negative | ESBL-producing Klebsiella sp/E.coli | 56 | 2 | 8 | 5 | 11 |
(35.9) | (3.1) | (22.2) | (62.5) | (34.4) | ||
Enterobacteriaceae* resistant to 3rd or 4th | 13 | 20 | 8 | – | 2 | |
generation cephalosporins | (8.3) | (30.8) | (22.2) | (6.3) | ||
CRGNB** | 25 | 22 | 11 | – | 5 | |
(16) | (33.9) | (30.6) | (15.6) | |||
Others | 33 | 8 | 8 | – | – | |
(21.1) | (12.3) | (22.2) | ||||
Total | 156 | 65 | 36 | 8 | 32 | |
(100) | (100) | (100) | (100) | (100) |
CLABSI: Catheter-associated bloodstream infection; VAP: Ventilator-associated pneumonia; CAUTI: Catheter-associated urinary tract infection; MRSA: Methicillin-resistant S. aureus, CRGNB: Carbapenem-resistant Gram-negative bacteria. #Methicillin-resistant S. aureus, ##Vancomycin-intermediate S. aureus, ###Vancomycin-resistant Enterococcus, *Enterobacter/Proteus/Morganella, Citrobacter, **CRGNB (Carbapenem-resistant Gram-negative bacteria, including Pseudomonas aeruginosa, Acinetobacter spp, and Enterobacteriaceae).
MRSA was the most reported CLABSI- and SSI-causing MDR bacteria, while ESBL producing Klebsiella sp and E.coli were the most reported CAUTI-causing MDR bacteria, and CRGNB were the most reported VAP-causing ones (Table 3).
TABLE 3: Infection-causing multi-drug-resistant (MDR) bacteria in children admitted to PICUs, according to the site, reported to the Rio de Janeiro state surveillance system (2014-2017)
CLABSI | VAP | CAUTI | SSI | ||
---|---|---|---|---|---|
(%) | (%) | (%) | (%) | ||
Gram- positive | MRSA# | 53 | 6 | 1 | 13 |
(31.4) | (6.9) | (1.9) | (56.5) | ||
VISA## | 4 | 1 | – | – | |
(2.4) | (1.1) | ||||
VRE### | 3 | – | – | – | |
(1.8) | |||||
Gram-negative | ESBL-producing Klebsiella sp and E.coli | 23 | 12 | 18 | 2 |
(13.6) | (13.8) | (34.6) | (8.7) | ||
Enterobacteriaceae* resistant to | 5 | 3 | 2 | 1 | |
3rd or 4th generation cephalosporins | (3.0) | (3.4) | (3.8) | (4.3) | |
CRGNB** | 51 | 50 | 17 | 5 | |
(30.2) | (57.5) | (32.7) | (21.7) | ||
Others | 30 | 15 | 14 | 2 | |
(17.8) | (17.2) | (26.9) | (8.7) | ||
Total | 169 | 87 | 52 | 23 | |
(100) | (100) | (100) | (100) |
CLABSI: Catheter-associated bloodstream infection; VAP: Ventilator-associated pneumonia; CAUTI: Catheter-associated urinary tract infection; MRSA: Methicillin-resistant S. aureus; CRGNB: Carbapenem-resistant Gram-negative bacteria. #Methicillin-resistant S. aureus, ##Vancomycin-intermediate S. aureus., ###Vancomycin-resistant Enterococcus, *Enterobacter/Proteus/Morganella, Citrobacter) **CRGNB ( Carbapenem-resistant Gram-negative bacteria- Includes Pseudomonas aeruginosa, Acinetobacter spp, and Enterobacteriaceae).
DISCUSSION
Monitoring the regional distribution of HAI-causing MDR bacteria is one of the most important cornerstones of infection control. It makes the recognition of prevalent agents and the establishment of governmental actions possible; thus, reducing their impact. Surveillance systems help in HAI tracking and prevention measure evaluation, provide data needed to identify critical issues, and contribute to HAI reduction14. Knowledge on colonized MDR bacteria patients contributes to the establishment of better infection control practices that prevent their spread in neonatal and pediatric intensive care units, and the definition of appropriate antimicrobial schemes in case of infection15–16.
Rio de Janeiro is the third most populous Brazilian state, with approximately 16.7 million inhabitants. Since 2013, efforts have been made to increase the reporting of colonization- and infection-associated MDR bacteria in adult and pediatric hospitals. In this study, an increase in the absolute number of colonizations and infections in NICUs and PICUs was observed when the annual values were compared with the mean value of the study period.
This survey revealed that ESBL producing Klebsiella sp and E.coli were the most reported colonization causing agents in both NICUs and PICUs, followed by MRSA. Critical agents such as carbapenem-resistant Gram-negative bacteria (CRGNB) were reported in 9.9% of neonates and 10.2% of children. Compared with this study, different rates of MDR bacteria colonization are reported in other studies. For example, Mammina et al. described a rate of 55.2% in neonates colonized by multidrug-resistant Gram-negative bacilli (MDRGN) during a one-year follow-up in a NICU in Italy17. However, Rybczmiska et al, in their study on the prevalence of ESBL bacteria colonization in a NICU in Sweden reported a rate of 1.77%18. Tfifha et al., reported a 4.9% rate of MDR acquisition during hospitalization in a PICU in Tunisia, and 7.84% at discharge, with ESBL-producing E. coli and K. pneumoniae being most frequently detected at discharge (52.7%)15.
In this study, the frequency of reported device-associated infection-causing MDR bacteria was described, and different patterns were identified among the most reported agents in neonates and children, based on the site of infection. For instance, ESBL-producing Klebsiella and E.coli were the most reported MDR bacteria, causing 35.9% of CLABSI in NICUs, while MRSA was the most reported agent, causing 31% of CLABSI in PICUs. Data from São Paulo State (Brazil) HAI surveillance system (2017) presented a global bloodstream infection-antibiotic resistance of 37.1% in NICUs, and 34% of CLABSI-causing Klebsiella pneumoniae isolates were 3rd generation cephalosporin-resistant, 50% of S. aureus were oxacillin-resistant and 75% of coagulase-negative staphylococci were oxacillin-resistant19. The Brazilian surveillance system, which compiled data from the entire country in 2016, reported 54% of oxacillin-resistant S. aureus in CLABSI from PICUs, 44.4% of carbapenems-resistant Acinetobacter spp, and 29.1% 3rd and 4th generation cephalosporins-resistant Klebsiella pneumoniae. When CLABSI data from NICUs were analyzed, it was observed that 53.6% of S. aureus were oxacillin-resistant, 37.8% of Klebsiella pneumoniae were 3rd and 4th generation cephalosporin-resistant, and 30.4% of E.coli were 3rd and 4th generation cephalosporin-resistant11.
Of all the reported MDR bacteria, carbapenem-resistant Gram-negative bacteria were the most reported VAP-causing agents in both NICUs and PICUs. In children and neonates, CRGNB represented more than half and a third of the reported VAP-causing MDR bacteria, respectively. These agents, which are considered critical by the World Health Organization (WHO), and for which there is an urgent need to development new antibiotics, usually have high levels of antimicrobial resistance (AMR), and are associated with high mortality20. When other healthcare institutions or surveillance systems were compared, the data available were limited, as described in a recent systematic review which investigated the main VAP-causing pathogens in Brazilian NICUs. The review reported that VAP incidence density ranged between 3.2 and 9.2 per 1,000 ventilator-days, and none of the papers reviewed in this study reported which pathogens were associated with VAP21. Lake et al., described pathogen distribution and antimicrobial resistance patterns for HAIs, including VAP reported to the National Healthcare Safety Network (NHSN) by pediatric locations from 2011-2014. Their study revealed that in NICUs, 13.4% and 9.1% of VAP-causing Pseudomonas aeruginosa and Acinetobacter spp, respectively, were carbapenem-resistant, while in PICUs, 16.3% and 12.5% of VAP-causing Pseudomonas aeruginosa and Klebsiella pneumoniae/oxytoca, respectively, were carbapenem-resistant9.
Interestingly, CRGNB were also the most reported and second most reported CAUTI causing agents among MDR bacteria in NICUs and PICUs, respectively. Like VAP, limited data is available in Brazilian surveillance systems on which CAUTI-causing agents are prevalent, and on the value of their resistance rates. Duenas et al. studied device-associated infections, including CAUTI, in a NICU and PICU in El Salvador during a 2-year follow-up period. Their study revealed that the main pathogens associated with CAUTI in children were Candida sp (36.3%), Pseudomonas sp (18.2%), and Klebsiella sp (18.2%), although no information about resistance was available22. United States’ data on bacterial resistance in HAI children and neonates showed a lower percentage of CAUTI-causing carbapenem-resistant bacteria in PICUs. For example, HAI-related bacteria resistance to carbapenems was 7.2%, 2.8%, 1.1%, and 0.7% for Pseudomonas aeruginosa, Enterobacter sp, Klebsiella pneumoniae/oxytoca, and E.coli, respectively, in PICUs, and the report described only data from PICUs and pediatric wards9.
In order to reinforce the role of surveillance, Rio de Janeiro State has compiled data on colonization-causing MDR bacteria and device-associated infection in NICUs and PICUs since 2014. The present study showed a high rate of MDR Gram-negative, specifically carbapenem-resistant bacteria, among all the MDR bacteria reported in NICUs and PICUs. MDR microorganism surveillance is one of the global action plans of WHO on antimicrobial resistance strategic objectives, published in 20151, after which, Brazil instituted a national prevention and control plan against antimicrobial resistance in 2018, including the development of scientific evidence in this area, and the expansion and development of a national MDR microorganisms-monitoring network23.
Compilation of data on MDR bacteria from different healthcare institutions with different rate of MDR presentation and the creation of regional benchmarking could be interpreted as limitations of this study. Secondly, retrospective data from an electronic web-based file was used; however, some healthcare units were unable to report MDR bacteria cases due to the lack of access to the system. Lastly, some healthcare institutions may have not entered data, due to MDR agent reporting incompetence.
In conclusion, MDR Gram-negative bacteria, including ESBL producing and carbapenem-resistant bacteria, were the most reported agents among all MDR bacteria cases that were reported to the Rio de Janeiro State surveillance system between 2014 and 2017, causing all device-associated infections in NICUs and PICUs, with the exception of CLABSI in PICUs. Efforts to increase MDR bacteria reporting to the surveillance system are necessary. This will contribute to the development of strategies to reduce the impact of HAI on neonates and children in the state of Rio de Janeiro.