Korean J healthc assoc Infect Control Prev 2023; 28(1): 113-125
Published online June 30, 2023 https://doi.org/10.14192/kjicp.2023.28.1.113
Copyright © Korean Society for Healthcare-associated infection Control and Prevention
Erdenetuya Bolormaa1*, Choryok Kang2*, Young June Choe3,4 , Joo Seon Heo3,5, Hannah Cho3
Department of Preventive Medicine, Korea University College of Medicine1, College of Nursing, Seoul National University2, Department of Pediatrics, Korea University Anam Hospital3, Allergy and Immunology Center, Korea University4, Institute of Nano, Regeneration, Reconstruction, Korea University College of Medicine, Korea University5, Seoul, Korea
Correspondence to: Young June Choe
E-mail: choey@korea.ac.kr
ORCID: https://orcid.org/0000-0003-2733-0715
*Equally contributed to the manuscript.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0).
Background: Catheter-related bloodstream infections (CRBSIs) are serious complications in neonatal intensive care units (NICUs). We aimed to assess the incidence of CRBSIs in NICUs worldwide and describe the causative organisms.
Methods: We searched PubMed, EMBASE, Cochrane, and KoreaMed databases. We included studies on CRBSIs in NICU settings with data on bacteremia. We performed a random-effects meta-analysis on CRBSI incidence in NICUs, stratified the data according to WHO regions. We compiled data on underlying organisms.
Results: Of the 692 studies identified, 71 published between 2011 and 2022 were considered eligible. The pooled incidence of CRBSI per 1000 catheter days in NICUs was 8.66 (95% confidence interval [CI], 7.19; 10.12). Stratifying by WHO regions, the CRBSI incidence per 1000 catheter days was 10.38 (95% CI, 3.86; 16.90) in the Eastern Mediterranean Region (EMR), 11.77 (95% CI, 9.20; 14.35) in the European Union Region (EUR), 5.94 (95% CI, 3.87; 8.00) in the Western Pacific Region (WPR), and 6.71 (95% CI, 4.39; 9.03) in the Region from the Americas (AMR). Of the 2887 bacterial strains, 73.4% (n=2118) were gram-positive bacteria, 18.9% (n=547) were gram-negative bacteria, and 7.8% (n=225) were fungi. Coagulasenegative Staphylococci (n=1380, 65.2%) were the most common pathogen among the grampositive types, followed by Staphylococcus aureus (n=318, 15%). Among the CRBSI gramnegative cultures, Klebsiella spp. (n=201, 36.7%) was the primary pathogen.
Conclusion: We found a substantial burden of CRBSIs in NICUs across the globe. Our findings highlight the need to improve the implementation of global and local strategies to reduce CRBSIs in NICUs.
Keywords: Neonate, Bacteremia, Catheters, Infection control, Systematic review
Central venous catheters are commonly used to administer medications and parenteral nutrition to vulnerable neonates in neonatal intensive care units (NICUs) [1]. A common and serious complication of central venous catheters is a catheter-related bloodstream infection (CRBSI), which is the most common cause of late-onset sepsis and has an estimated mortality rate of 70% in infants [2]. Neonates are highly vulnerable to CRBSIs in NICUs; however, incidence estimates are lacking in many countries.
A previous systematic review and meta-analysis investigated the incidence of neonatal sepsis [3]; however, no systematic review on the global incidence of CRBSIs limited to NICU settings has been reported to date. The study sought to review neonatal sepsis and mortality across low- and middle-income countries; however, this is not specific to CRBSIs in the NICU [3]. Furthermore, no regional comparative study has investigated the incidence of CRBSIs in NICUs.
We conducted a systematic literature review to assess the incidence of CRBSIs in NICUs worldwide and describe the causative organisms. We aimed to assess the global incidence of CRBSI, particularly in NICUs, and compile data on causative pathogens.
We searched PubMed, EMBASE, Cochrane, and KoreaMed databases using the following keywords: “catheter-related infection”, “CVC infection”, “CVC-related infection”, “CVC associated infection”, “central line infection”, “central line related infection”, “central line-associated infection”, “bacteremia”, “bloodstream infection”, “neonatal intensive care”, “NICU”, “infant”, “neonate”, “newborn”, “newly born infant”, “neonatal infant”, “premature infant”, “preterm infant”, “low birth weight infant”, “LBW infant”, and “CRBSI”. We combined these terms with “AND” or “OR” when searching the databases. The search was performed on December 9, 2022.
Studies were reviewed by their titles and abstracts in the first screening and by full-textarticles in the second screening. The inclusion criteria were as follows: (1) the research participants included patients with CRBSIs, (2) all patients must have been admitted to the NICU, (3) the research participants had received no prior interventions, (4) the research was about microorganisms, and (5) the study was published in English. The exclusion criteria were as follows: (1) the study was duplicated, (2) the data could not be extracted or converted for useful data, (3) the studies were case reports, roundtable meeting reports, conference reports, or reviews, (4) the study was published in languages other than English, and (5) the results were incomplete.
The information about the first author, published year, research time, country, data collection method, total patients in NICUs, total patients with CRBSIs, total catheter days, and total distribution of species and types of microorganisms was extracted by the review investigator with Microsoft Office Excel 2010.
The proportion of extracted data, the proportion of pathogens, and the pooled incidence and its 95% confidence intervals were analyzed using a meta package of R 4.2.2. A random effects model was chosen based on the heterogeneity and significance tests (
A total of 692 studies were first screened by their titles and abstracts. and 331 were selected for the next full-text screening. Among those, 97 studies did not involve patients with CRBSIs in NICUs, 9 recruited intervention participants, 2 studies were reviews, 9 were written in French and Chinese languages, and 143 did not have useful data for our meta-analysis of CRBSI incidence. Finally, 71 studies [4-74] were eligible for our analysis. We depicted the flowchart of the selection of surveys in Fig. 1.
Table 1 describes the selected studies on CRBSIs in NICUs.
Table 1 . The characteristics of studies included in this systematic review
Ref No. | Study | No. of subjects | Study design | Country | Study year | Study population | CLABSI prevention | Incidence/1000 catheter-days |
---|---|---|---|---|---|---|---|---|
[4] | Al-Mousa et al. | 671 | Prospective cohort | Kuwait | 2013-2015 | Neonatal patients | None | 15.3 |
[5] | Almeida et al.* | 1194 | Retrospective | Portugal | 2007-2010 | Newborn infants | Preventive bundle | 14.1 |
[6] | Arnts et al.* | 45 | Prospective observational | Netherland | 2009-2010 | Newborn infants | Preventive bundle | 12.9 |
[7] | Bannatyne et al.* | 406 | Retrospective cohort | Australia | 2011-2013 | Newborn infants | Preventive bundle | 8.8 |
[8] | Bierlaire et al.* | 140 | Prospective | Belgium | 2019 | Neonates | Preventive bundle | 8.4 |
[9] | Blanchard et al. | Retrospective cohort | Canada | 2007-2011 | Neonatal patients | None | 4 | |
[10] | Bolat et al. | 569 | Prospective, cohort | Turkey | 2009-2011 | Neonatal patients | None | 3.64 |
[11] | Boutaric et al.* | 111 | Prospective | France | 2004-2006 | Premature infants | Preventive protocol | 16 |
[12] | Bunni et al.* | 311 | Retrospective | UK | 2009 | Neonates | Preventive bundle | 22.4 |
[13] | Cabrera et al. | 167 | Prospective | Peru | 2017-2018 | Neonates | None | 8 |
[14] | Callejas et al. | 689 | Retrospective | Canada | 2010-2013 | Neonates | None | 5.6 |
[15] | Chandonnet et al.* | Prospective | USA | 2011 | Neonatal patients | Preventive bundle | 2.6 | |
[16] | Cheng et al. | 123 | Retrospective cohort | China | 2011-2012 | Neonates | None | 4.99 |
[17] | Cheong et al. | 39 | Retrospective | Japan | 2013 | VLBW infants | None | 3.57 |
[18] | Cleves et al. | 1246 | Retrospective, quasi-experimental | USA | 2012-2014 | Neonates | Chlorhexidine baths | 8.64 |
[19] | Dumpa et al.* | 68 | Retrospective review | USA | 2009-2010 | Neonatal patients | Preventive bundle | 4.4 |
[20] | Erdei et al.* | Prospective | USA | 2009 | Newborn infants | Preventive bundle | 4.1 | |
[21] | Ereno et al. | 107 | Retrospective | Singapore | Neonatal patients | None | 5.9 | |
[22] | Flidel-Rimon et al.* | 141 | Prospective | Israel | 2011-2012 | Infants | Preventive bundle | 15.2 |
[23] | Fontela et al. | Retrospective dynamic cohort | Australia | 2003-2009 | Neonatal patients | None | 4.4 | |
[24] | Freeman et al.* | 285 | Retrospective | USA | 2005-2012 | Neonatal patients | Prevention protocol | 1.69 |
[25] | Freitas et al. | 1560 | Prospective cohort | Brazil | 2014-2016 | Neonates | None | 18.6 |
[26] | Gadallah et al. | 434 | Prospective cohort | Egypt | 2012 | Neonates | None | 158.3 |
[27] | Gerver et al. | Retrospective | UK | 2016-2017 | Neonates | None | 1.5 | |
[28] | Greenhalgh et al. | 176 | Retrospective cohort | Australia | 2012 | Neonates | None | 11.5 |
[29] | Hei et al. | 131 | Prospective | China | 2008-2011 | Neonatal patients | None | 13.7 |
[30] | Helder et al.* | 537 | Prospective, observational | Netherland | 2014-2016 | Infants | Antiseptic protocol | 3.1 |
[31] | Hocevar et al. | Retrospective | USA | 2006-2008 | Neonates | None | 3.9 | |
[32] | Holzmann-Pazgal et al.* | Retrospective | USA | 2006-2008 | Neonates | Line team | 11.6 | |
[33] | Hussain et al. | 301 | Prospective | Pakistan | 2016 | Neonatal patients | Preventive bundle | 17.1 |
[34] | Hussain et al. | 2046 | Retrospective | Pakistan | 2011-2015 | Neonatal patients | None | 8.9 |
[35] | Jansen et al. | 180 | Retrospective cohort | Netherland | 2015-2019 | Preterm neonates | None | 14 |
[36] | Jansen et al. | 891 | Retrospective cohort | Netherland | 2012-2020 | Preterm neonates | None | 13.4 |
[37] | Jeong et al.* | 326 | Retrospective | Korea | 2011-2013 | Neonatal patients | Preventive bundle | 6.6 |
[38] | Kim et al. | Retrospective review | Korea | 2016-2020 | Infants | None | 2.85 | |
[39] | Kinoshita et al. | 2383 | Prospective observational | Japan | 2014-2017 | VLBW infants | None | 2.1 |
[40] | Kleinlugtenbeld et al.* | 75 | Prospective | Netherland | 2007 | Premature newborn | Preventive bundle | 20.1 |
[41] | Kourkouni et al. | Prospective | Greece | Neonatal patients | None | 6.58 | ||
[42] | Kulali et al.* | 70 | Prospective cohort | Turkey | 2016-2017 | Neonatal patients | Preventive bundle | 12.4 |
[43] | Leblebicioglu et al. | 3430 | Prospective | Turkey | 2003-2012 | Neonatal patients | None | 21 |
[44] | Leistner et al. | 5586 | Prospective cohort | Germany | 2008-2009 | VLBW infants | None | 8.3 |
[45] | Leveillee et al. | 1577 | Retrospective cohort | Canada | 2011-2016 | Neonates | None | 8.4 |
[46] | Milstone et al. | 3967 | Retrospective cohort | USA | 2005-2010 | Neonates | None | 1.66 |
[47] | Mohamed Cassim et al.* | 350 | UK | 2010-2011 | Newborn infants | Preventive bundle | 4.3 | |
[48] | Nercelles et al. | 4704 | Prospective | Chile | 2005-2011 | Newborn infants | None | 0.9 |
[49] | Nielsen et al. | 382 | Retrospective | Denmark | 2019-2020 | Neonatal patients | None | 13.41 |
[50] | Oh et al. | 429 | Retrospective | Korea | 2017 | Infants | Preventive bundle | 1.89 |
[51] | Patrick et al.* | Prospective cohort | USA | 2007-2012 | Neonatal patients | None | 2.1 | |
[52] | Pavcnik-Arnol et al. | Prospective cohort | Slovenia | 2011-2012 | Neonatal patients | None | 5.5 | |
[53] | Pharande et al.* | 13731 | Prospective | Australia | 2002-2016 | Newborn infants | Preventive bundle | 12.04 |
[54] | Piazza et al. | Retrospective | USA | 2011 | Neonatal patients | None | 1.333 | |
[55] | Ponnusamy et al. | 189 | Prospective observational | UK | 2009-2010 | Infants | None | 16.9 |
[56] | Rallis et al.* | 94 | Prospective | Greece | 2012 | Neonates | Preventive bundle | 12 |
[57] | Resende et al.* | 551 | Prospective | Brazil | 2010-2011 | Infants | Preventive bundle | 23 |
[58] | Rosenthal et al.* | 2009 | Prospective surveillance | El Salvador, Mexico, Philippines, Tunisia | 2003 | Neonatal patients | Preventive bundle | 21.4 |
[59] | Salm et al.* | 3028 | Prospective cohort | Germany | 2007-2009 | VLBW infants | Preventive bundle | 13.47 |
[60] | Sanderson et al. | 4248 | Prospective | Australia | 2007-2009 | Infants | None | 10.6 |
[61] | Shalabi et al. | 540 | Retrospective matched cohort | Canada | 2010-2013 | Infants | None | 8.5 |
[62] | Shepherd et al.* | USA | 2003-2006 | Infants | Preventive bundle | 6 | ||
[63] | Sinha et al.* | 152 | Retrospective | UK | 2007 | Preterm neonates | Preventive bundle | 26.5 |
[64] | Soares et al. | 251 | Retrospective cohort | Portugal | 2014-2016 | Neonatal patients | None | 12.4 |
[65] | Steiner et al.* | 526 | Prospective | Germany | 2010-2012 | VLBW infants | Preventive bundle | 8.96 |
[66] | Taylor et al. | 83 | Retrospective, quasi-experimental | Australia | 2013-2017 | Infants | None | 13.8 |
[67] | Ting et al.* | Retrospective observational | Canada | 2007-2008 | Neonates | Preventive bundle | 7.9 | |
[68] | Wen et al. | 301 | Prospective | China | 2010-2014 | Premature infants | None | 1.9 |
[69] | Wilder et al.* | USA | 2011 | Neonatal patients | Preventive bundle | 3.9 | ||
[70] | Worth et al. | Prospective | Australia | 2008-2016 | Neonatal patients | None | 2.2 | |
[71] | Yalaz et al. | 1200 | Prospective | Turkey | 2008-2010 | Newborn infants | None | 4.1 |
[72] | Yumani et al. | 369 | Retrospective | Netherland | 2007 | Neonatal patients | None | 18.1 |
[73] | Zachariah et al. | Cross-sectional | USA | 2011 | Neonatal patients | None | 1.52 | |
[74] | Zhou et al. | 29 | Prospective | China | 2008-2010 | Newborns | Preventive bundle | 16.7 |
*Estimated only pre-intervention period CLABSI rate.
Abbreviation: VLBW, very low birth weight.
The 71 eligible studies were published from 2011-2022, mainly concentrated in 2012-2016. The surveys were conducted between 2002 and 2020. Dividing the studies by WHO regions, 22 (30.9%) were from the Region from the Americas (AMR) [9,13-15,18-20,24,25,31,32,45,46, 48,51,54,57,61,62,67,69,73], 5 (7.0%) from the Eastern Mediterranean Region (EMR) [4,26,33,34,58], 27 (38.0%) from the European Union Region (EUR) [5,6,8, 10-12,22,27,30,35,36,40-44,47,49,52,55,56,59,63-65,71,72], and 17 (23.9%) from the Western Pacific Region (WPR) [7,16,17,21,23,28,29,37-39,50,53,60,66,68, 70,74].
Most studies were from the United States (n=13, 18.3%) [15,18-20,24,31,32,46,51,54,62,69,73], Australia (n=7, 9.9%) [7,23,28,53,60,66,70], and the Netherlands (n=6, 8.5%) [6,30,35,36,40,72]. Regarding the methodology, 50.7% (n=36) were retrospective and 45.1% (n=32) were prospective studies. We included 63 082 patients in NICUs, except for 17 studies that did not provide information on the total number of patients with catheters.
We estimated the pooled incidence of CRBSI per 1000 catheter days in NICUs by dividing the regions into subgroups. The CRBSI incidence per 1000 catheter days was 10.38 in the EMR (95% CI, 3.86; 16.90), 11.77 (95% CI, 9.20; 14.35) in the EUR, 5.94 (95% CI, 3.87; 8.00) in the WPR, and 6.71 (95% CI, 4.39; 9.03) in the AMR, and the total weighted CRBSI incidence per 1000 catheter days was 8.66 (95% CI, 7.19; 10.12) (Fig. 2).
Fig. 3 shows the trend of CRBSI per 1000 catheter days by an identified period of surveillance. The incidence of CRBSI per 1000 catheter days was 0.0-26.5, 1.9-23, and 1.5-17.1 in 2006-2010, 2011-2015, and 2016-2020, respectively.
A total of 2887 bacterial strains were isolated from CRBSI samples. Among these, 73.4% (n=2118) were gram-positive bacteria, 18.9% (n=547) were gram-negative bacteria, and 7.8% (n=225) were fungi. Coagulase-negative Staphylococci (n=1380, 65.2%) was the most common pathogen among the gram-positive type, followed by
Table 2 . The proportion of pathogen related to catheter-related bloodstream infection in neonatal intensive care unit (2006-2020)
Total pathogens (n=2887) | Frequency (%) |
---|---|
Gram-positive (n=2118) | |
Coagulase-negative Staphylococci | 1380 (65.2) |
88 (4.2) | |
14 (0.7) | |
333 (15.7) | |
Methicillin-resistant | 14 (0.7) |
Methicillin-sensitive | 10 (0.5) |
166 (7.8) | |
64 (3) | |
19 (0.9) | |
11 (0.5) | |
Other gram-positive | 19 (0.9) |
Gram-negative (n=547) | |
201 (36.7) | |
37 (6.8) | |
96 (17.6) | |
65 (11.9) | |
6 (1.1) | |
32 (5.9) | |
24 (4.4) | |
19 (3.5) | |
5 (0.9) | |
3 (0.5) | |
2 (0.4) | |
2 (0.4) | |
2 (0.4) | |
Other gram-negative | 53 (9.7) |
Fungi (n=225) | |
170 (75.6) | |
33 (14.7) | |
10 (4.4) | |
5 (2.2) | |
3 (1.3) | |
2 (0.9) | |
2 (0.9) |
Gram-positive species were the most common pathogen type among CRBSI incidences in all three regions in our study, with proportions of 70% in the AMR, 76% in the WPR, and 84% in the EUR. Fig 4 describes the proportions of pathogen types among the subgroups of WHO regions. Among the EMR region surveys, no survey isolated pathogenic species. Approximately 20%, 21%, and 14% of the strains were gram-negative in the AMR, WPR, and EUR, respectively. Fungi was isolated in only 2% in EUR, 10% in AMR, and 3% in WPR.
We analyzed a total of 71 studies and showed a substantial burden of CRBSIs in NICUs globally; however, our review was limited by a vast difference in terms of incidence rate, necessitating a standardized investigative method to report CRBSIs in NICUs. According to the National Healthcare Safety Network in the United States, only CLABSI in children ≤1 year is defined, which may not be suitable for neonates due to differences in the symptoms of infection [75]. This should motivate global researchers to define local NICU CRBSI definitions according to the standardized recommendations and sustainably implement such preventive measures. In this context, a modified case definition for CRBSI in NICU settings should be adapted from the previously defined “catheter-related bloodstream infection” or “central line-associated bloodstream infection” [76].
This systematic literature review is the first to investigate the global incidence of CRBSIs in NICUs. We estimated the CRBSI incidence with stratification according to WHO regions and identified regional differences in CRBSI incidence. In this study, incidence estimates were higher in the EMR and EUR than in the WPR and AMR. This finding indicates that the need to reduce CRBSIs in NICUs is greater in the EMR and EUR. Data from the African and Southeast Asian regions were not included in this study, which might lead to knowledge gaps on the global incidence of CRBSIs in NICUs. The incidence of neonatal sepsis is reported to be very high in the African region; however, the majority of hospital-wide and ICU-based studies have been conducted in high-income regions such as the European and American WHO regions [77]. Therefore, further studies are required to investigate the data from the African and Southeast Asian regions.
Furthermore, we found a downward trend in the incidence of CRBSIs in NICUs across countries. This may be explained by the adoption of CRBSI prevention bundles at multiple sites; however, this could not be determined from the current dataset. We propose a longitudinal analysis in defined clinical settings to investigate the role of prevention bundles in the incidence of CRBSIs in NICUs at different times.
In our study, the most common causative pathogens of CRBSIs in NICUs were coagulase-negative Staphylococci,
Reducing neonatal mortality is an important component of the third Sustainable Development Goal. It is essential to understand the wide variability in neonatal health outcomes, particularly in NICUs across the globe. A recent systematic review showed that West and Central Africa and South Asia had the highest neonatal mortality rates in 2017, despite improvements from the 1990s [78]. A vast difference in neonatal mortality due to infectious causes between high- and low-income countries and regions is an important issue since the regionalization of neonatal healthcare is emphasized [79]. In this context, our study showed that neonatal CRBSI incidence was variable across countries, particularly in different settings. Higher incidences were observed in the Eastern Mediterranean and European regions compared to those of the Western Pacific and American regions. This difference may be explained by the access to facilities for newborns, as previously described [80].
Our study was limited by the variability among individual studies, resulting in heterogeneity of the synthesized data, which required careful interpretation. Despite our broad search strategy with a focus on CRBSIs in NICUs, data from the African and Southeast Asian regions were not included in our study. This may be explained by a lack of epidemiological data, deficiencies in healthcare organizations and resources, and institutional obstacles to delivering critical care in the resource-limited settings of low-income countries. A recent systematic review showed that West and Central Africa and South Asia had the highest neonatal mortality rates in 2017, despiteimprovements from the 1990s [81]. Reducing neonatal mortality is an important component of the third Sustainable Development Goal. It is essential to understand the wide variability in neonatal health outcomes, particularly in NICUs across the globe.
We found a variable incidence of CRBSIs in NICUs globally, with a downward trend over the past 15 years; however, the substantialdisease burden remains among newborns. Our findings highlight the need to improve the implementation of global and local strategies to reduce CRBSIs in NICUs. Future research is required to address the knowledge gaps identified by our study.
This study was supported by the Korean Society for Healthcare-Associated Infection Control and Prevention in 2021.
The authors have no potential conflict of interest to disclose.
Song Yi An, Woong Jung, Chang Min Lee, Sung Hyuk Park, Hyun Kyung Park, Myung Chun...
2017; 22(2): 43-55 https://doi.org/10.14192/kjicp.2017.22.2.43Song Yi An, Woong Jung, Chang Min Lee, Sung Hyuk Park, Hyun Kyung Park, Myung Chun...
2024; 29(2): 128-136 https://doi.org/10.14192/kjicp.2024.29.2.128Song Yi An, Woong Jung, Chang Min Lee, Sung Hyuk Park, Hyun Kyung Park, Myung Chun...
2024; 29(2): 103-109 https://doi.org/10.14192/kjicp.2024.29.2.103