As reported by the International Nosocomial Infection Control Consortium (INICC) in pooled studies, rates of central-line–associated bloodstream infection (CLABSI) have been significantly higher in low- and middle-income countries (LMICs) than in high-income countries.Reference Rosenthal, Maki and Salomao1–Reference Rosenthal3 According to a review, they ranged from 1.6 to 44.6 CLABSIs per 1,000 central-line (CL) days in adult and pediatric intensive care units (ICUs) and from 2.6 to 60.0 CLABSIs per 1,000 CL days in neonatal NICUs (NICUs).Reference Rosenthal3 CLABSI rates significantly increased in ICUs of LMICs during the COVID-19 pandemic.Reference Rosenthal, Myatra and Divatia4
Multiple logistic regression identified the acquisition of a CLABSI as an independent risk factor associated with ICU all-cause mortality (adjusted odds ratio [aOR], 1.84; 95% confidence interval [CI], 1.73–1.95; P < .0001).Reference Rosenthal, Yin and Lu5 This association was also demonstrated in 9 Asian countries (aOR, 2.36; 95% CI, 2.14–2.61; P < .0001)Reference Rosenthal, Jin and Rodrigues6 and 10 Middle Eastern countries (aOR, 1.49; 95% CI, 1.33–1.66; P < .00001).Reference Rosenthal, Jin and Memish7
CLABSIs are associated with higher mortality by 12%–25%8 and extra costs.Reference Stone, Braccia and Larson9–Reference Tarricone, Torbica, Franzetti and Rosenthal12 The INICC reported that mortality in ICU patients without any healthcare-associated infection (HAI) is 17.1%, with CLABSI the mortality rate is 48.2%, and mortality is 63.4% with a CLABSI plus catheter-associated urinary tract infection (CAUTI) plus ventilator-associated pneumonia (VAP).Reference Rosenthal, Duszynska and Ider2
Previous studies have identified the following variables as CLABSI risk factors: body mass index >40 kg/mReference Rosenthal, Duszynska and Ider2,Reference Buetti, Marschall and Drees13 multiple CLs,Reference Buetti, Marschall and Drees13,Reference Almuneef, Memish, Balkhy, Hijazi, Cunningham and Rate14 multilumen catheters,Reference Buetti, Marschall and Drees13 femoral site,Reference Lorente, Henry, Martin, Jimenez and Mora15,Reference O’Horo, Maki, Krupp and Safdar16 guidewire exchange,Reference Almuneef, Memish, Balkhy, Hijazi, Cunningham and Rate14 heavy microbial colonization at insertion site or catheter hub,Reference Buetti, Marschall and Drees13 indwelling time,Reference Buetti, Marschall and Drees13 prolonged hospitalization before catheterization,Reference Buetti, Marschall and Drees13 neutropenia,Reference Buetti, Marschall and Drees13 total parenteral nutrition,Reference Buetti, Marschall and Drees13,Reference Almuneef, Memish, Balkhy, Hijazi, Cunningham and Rate14 patient cared for by a floating nurse,Reference Alonso-Echanove, Edwards and Richards17 transfusion of blood products,Reference Buetti, Marschall and Drees13 prematurity,Reference Buetti, Marschall and Drees13 reduced ICU nurse-to-patient ratio,Reference Buetti, Marschall and Drees13 substandard CL care,Reference Buetti, Marschall and Drees13 and few others.
However, no study has analyzed multiple countries or different types of vascular catheters simultaneously to identify CLABSI risk factors in ICUs. Also, no study has been conducted prospectively over 24 years. Furthermore, no study has analyzed all the following 11 variables simultaneously and their association with CLABSI: sex, age, length of stay (LOS), CL days before acquisition of CLABSI, CL device utilization (DU) ratio as a marker of severity of illness of patients, different types of vascular catheters, tracheostomy use, hospitalization type, ICU type, facility ownership, and World Bank country classifications by income level.
The objective of this study was to report CLABSI rates per country, per continent, per type of ICU, per facility ownership, per income level (according to the World Bank), and per year. We also analyzed whether these 11 variables were CLABSI risk factors, and we sought to identify the safest type of CL.
Methods
Study population and design
This multinational, multicenter, cohort, prospective study included patients admitted to 728 ICUs of 286 hospitals in 147 cities in 41 countries of Africa, Asia, Eastern Europe, Latin America, and the Middle East across 24 years, between July 1, 1998, and February 12, 2022.
INICC surveillance online system
According to standard Centers for Disease Control and Prevention National Healthcare Safety Network (CDC NHSN) methods, HAI denominators are device days collected from all patients as pooled data, without specifying each patient’s characteristics or the number of device days related to such patients.Reference Dudeck, Edwards and Allen-Bridson18 INICC HAI surveillance is carried out using an online platform, the ISOS, which includes CDC NHSN criteria and methods.Reference Dudeck, Edwards and Allen-Bridson18 The ISOS also adds the collection of patient-specific data on all patients, with and without HAI.Reference Rosenthal19 Data were collected for all patients admitted to the ICU, which allowed matching by various characteristics and facilitated the estimation of CLABSI risk factors.
Prospective cohort surveillance of healthcare-associated infections
The data were collected on each patient at the time of their ICU admission. From admission to discharge, infection prevention professionals (IPPs) went to the bedside of each patient daily. All patients admitted to an ICU were prospectively included in this investigation, and their data were collected using the INICC surveillance online system (ISOS). Each IPP used a tablet at the bedside of each hospitalized patient in the ICU, logged into the ISOS, and uploaded the patient data in real time.Reference Rosenthal19 At the time the patient was admitted, this information included details about the setting, country, city, admission date, and ICU type, as well as patient data, sex, age, hospitalization type, and used of invasive devices. Each IPP uploaded information about invasive devices and positive cultures until patient discharge.Reference Rosenthal19 In patients with signs or symptoms of infection, an infectious diseases specialist approached the patient to determine the presence of HAI.Reference Rosenthal19
Each participating hospital had a microbiology laboratory that identified microorganism profiles and bacterial resistance. Furthermore, 35 patients with missing data regarding age and/or sex (0.01% of the sample) were excluded from this analysis. This study was approved by the institutional review boards of the hospitals involved. All patient and hospital identifiers were kept confidential.
Study definitions
Healthcare-associated infection
Healthcare-associated infection (HAI) definitions used during surveillance were those published by Centers for Disease Control and Prevention (CDC) in 199120 and all subsequent updates.Reference Dudeck, Edwards and Allen-Bridson18 Over the 24 years of this study, all IPPs of the participant hospitals have applied the current and updated CDC definition of HAI. That is, whenever the CDC updated their definition, IPPs began using the new updated definitions.Reference Dudeck, Edwards and Allen-Bridson18,Reference Emori, Culver and Horan20
Central line
A central line (CL) was defined as an intravascular catheter that terminated at or close to the heart or in one of the great vessels and was used for infusion, withdrawal of blood, or hemodynamic monitoring. The following are considered great vessels: aorta, pulmonary artery, superior or inferior vena cava, brachiocephalic veins, internal jugular veins, subclavian veins, external iliac veins, common iliac veins, femoral veins, in neonates, the umbilical artery or vein.Reference Dudeck, Edwards and Allen-Bridson18
Primary bloodstream infection
Primary bloodstream infection was defined as a laboratory-confirmed bloodstream infection (LCBI) that was not secondary to an infection at another body site.Reference Dudeck, Edwards and Allen-Bridson18
Central-line–associated bloodstream infection
Central-line–associated bloodstream infection (CLABSI) was defined as an LCBI in which an eligible BSI organism was identified and an eligible CL was present on or the day before.Reference Dudeck, Edwards and Allen-Bridson18
Laboratory-confirmed bloodstream infection 1
This term was used in a patient of any age who had a recognized bacterial or fungal pathogen not included on the NHSN common commensal list. This pathogen was identified from 1 or more blood specimens obtained by a culture or identified to the genus or species level by non–culture-based microbiologic testing methods. In addition, organism(s) identified in the blood were not related to an infection at another site.Reference Dudeck, Edwards and Allen-Bridson18
Laboratory-confirmed bloodstream infection 2
This term was used in a patient of any age who had at least 1 of the following signs or symptoms: fever (>38.0°C), chills, or hypotension. Also, the organism(s) identified in the blood are not related to an infection at another site, and the same NHSN common commensal is identified by culture from 2 or more blood specimens collected on separate occasions.Reference Dudeck, Edwards and Allen-Bridson18
Common commensal
Common commensal organisms included but were not limited to diphtheroids (Corynebacterium spp not C. diphtheria), Bacillus spp (not B. anthracis), Propionibacterium spp, coagulase-negative staphylococci (including Staphylococcus epidermidis), viridans-group streptococci, Aerococcus spp, Micrococcus spp, and Rhodococcus spp.Reference Dudeck, Edwards and Allen-Bridson18
Central-line–device utilization ratio
The central-line–device utilization ratio (CL-DU) was calculated as a ratio of CL days to patient days for each location type. As such, the CL-DU of a location measures the use of invasive devices and constitutes an extrinsic CLABSI risk factor. The CL-DU ratio also serve as a marker for the severity of illness of patients which is an intrinsic HAI risk factor.Reference Dudeck, Edwards and Allen-Bridson18
World Bank country classification by income level
The World Bank assigns the world’s economies to 4 income groups: low, lower-middle, upper-middle, and high. The classifications are based on gross national income (GNI) per capita in the current US dollars. Low-income countries are those with GNI < US$1,045. Lower-middle income countries are those with GNI from US$1,046 to US$4,095. Upper-middle income countries are those with GNI from US$4,096 to US$12,695. High-income countries are those with GNI > US$12,695.21 The inclusion of high-income countries allowed us to compare the risk factors for CLABSI among LMICs with those of high-income countries and identify if the income of the country is independently associated as a risk factor for CLABSI.
Facility or institution ownership type
Publicly owned facilities are owned or controlled by a governmental unit or another public corporation, where control is defined as the ability to determine the general corporate policy. Not-for-profit, privately owned facilities are legal or social entities created for the purpose of producing goods and services, whose status does not permit them to be a source of income, profit or other financial gains for the unit(s) that establish, control, or finance them. For-profit, privately owned facilities are legal entities set up for the purpose of producing goods and services and can generate a profit or other financial gains for their owners.22
Statistical analyses
For risk factor analysis, we conducted a case–control study nested in a prospective cohort study. Patients with and without CLABSI were compared using multiple logistic regression. Statistically significant variables were independently associated with an increased risk for CLABSI. We used was the Wald test, and statistical significance was set at .05. Calculated from the outputs of multiple logistic regression, adjusted odds ratios (aORs) and the corresponding 95% CIs of statistically significant variables were also reported.
We analyzed following 11 variables and their association with CLABSI: (1) age; (2) male or female sex; (3) LOS before acquiring a CLABSI; (4) CL days before acquisition of CLABSI; (5) CL-DU ratio as a marker of severity of illness of patient; (6) different types and insertion sites of vascular catheters (ie, internal jugular, femoral, arterial, subclavian, temporary catheter for hemodialysis, peripherally inserted central catheter or PICC); (7) tracheostomy use; (8) medical or surgical hospitalization type; (9) ICU type (ie, medical-surgical, medical, pediatric, surgical, coronary, neurosurgical, cardio-thoracic, neurologic, trauma, oncology pediatric, or oncology adult); (10) facility ownership (publicly owned, not-for-profit privately owned, for-profit privately owned, and teaching hospital)22 ; and (11) income level per country according to the World Bank (ie, low, lower–middle, upper–middle, or high).21 The evaluated outcome was the acquisition of CLABSI according to the CDC NHSN definitions.Reference Dudeck, Edwards and Allen-Bridson18
For analyses of CLABSI risk factors, we used data from 35 countries: Argentina, Bahrain, Brazil, Bulgaria, China, Colombia, Costa Rica, Cuba, Dominican Republic, Ecuador, Egypt, El Salvador, India, Jordan, Kosovo, Kuwait, Lebanon, Malaysia, Mexico, Mongolia, Morocco, Nepal, Pakistan, Panama, Peru, Philippines, Poland, Romania, Saudi Arabia, Slovakia, Sri Lanka, Thailand, Turkey, United Arab Emirates, and Vietnam. All of these countries collected all 11 independent variables of interest: sex, age, LOS, CL days, CL-DU ratio, different types and insertion sites of vascular catheters, tracheostomy use, hospitalization type, ICU type, facility ownership, and income level per country.
For estimation of CLABSI rates, we used data from all 41 countries including the 35 countries listed above and following 6 countries that collected only CL days and CLABSI events: Greece, Macedonia, Papua New Guinea, Russia, Serbia, Tunisia.
All statistical analyses were performed using R version 4.1.3 software (R Foundation for Statistical Computing, Vienna, Austria).
Results
From July 1, 1998, to February 12, 2022, over 24 years, a multinational, multicenter, cohort, prospective surveillance study of CLABSIs was conducted in 728 ICUs of 286 hospitals in 147 cities in 41 countries from Africa, Asia, Eastern Europe, Latin America, and the Middle East, participating in INICC. Patients admitted to Asian facilities represent 53.39% of the sample, followed by patients in Latin America (22.79%), the Middle East (21.11%), and Eastern Europe (2.71%).
In this cohort study, the length of participation of hospitals varied from 1.1 to 226.07 months (mean, 38.47; SD, 42.62). Table 1 shows data pertaining to setting and patient characteristics. Table 2 and Fig. 1 show CLABSI rates per 1,000 CL days per country and per region. Table 3 shows CLABSI rates stratified by ICU type, income level according to the World Bank, and facility ownership. Fig. 2 shows CLABSI rates stratified by year.
Fig. 1. Rate of CLABSI per 1,000 central line days, stratified by country.
Table 1. Setting and Patient Characteristicsa
Note. ICU, intensive care unit; CL, central-line; DU, device utilization; LOS, length of stay; CLABSI, central-line–associated bloodstream infection; SD, standard deviation.
a Data collected from July 1, 1998, to February 12, 2022, over 24 years.
b Data are no. (%) unless otherwise specified.
c Peripherally inserted central catheter.
Table 2. Central-Line–Associated Bloodstream Infections Rates Stratified per Country and per Region
Note. CL, central line; DU, device utilization; CLABSI, central-line–associated bloodstream infection; CI, confidence interval.
a Rate of CLABSI per 1,000 CL days.
Table 3. Central-Line–Associated Bloodstream Infections Rates Stratified per ICU Type, According to World Bank Country Classifications by Income Level and Facility Ownership Type
Note.
ICU, intensive care unit; CI, confidence interval.
a ICUs are listed in order of the highest to lowest central-line–associated bloodstream infection (CLABSI) rate.
Using multiple logistic regression, the following variables were identified as statistically significantly associated with CLABSI (Table 4): LOS, risk increasing 3% daily; number of CL days, risk increasing 4% per CL day; surgical hospitalization; tracheostomy; hospitalization at a publicly owned facility or at a teaching hospital; and hospitalization in a middle-income country. The ICU type with highest risk was adult oncology, followed by pediatric oncology, pediatric, and medical. The CL types with the highest risk were internal-jugular and femoral. PICC had the lowest risk for CLABSI.
Table 4. Multiple Logistic Regression Analysis of Risk Factors for Central-Line–Associated Bloodstream Infections
Note. CI, confidence interval; CL, central line; DU, device utilization; ICU, intensive care unit; LOS, length of stay; CLABSI, central-line–associated bloodstream infection; aOR, adjusted odds ratio.
a Peripherally inserted central catheter.
Discussion
Pooled rates of CLABSI in our study of 4.82 CLABSI per 1,000 CL days were similar to the pooled CLABSI rates reported by the INICC of 5.30 CLABSIs per 1,000 CL days.Reference Rosenthal, Duszynska and Ider2 Our present study, which shows CLABSI rates stratified per country and per region, contributes to the identification of countries and regions with higher and lower CLABSI rates (See Fig. 1). On the other hand, the pooled rate of CLABSI in our present study was significantly higher than that reported by the CDC NHSN of 0.8 CLABSIs per 1,000 CL days.Reference Dudeck, Edwards and Allen-Bridson18 In this study, we detected a trend of significant reduction in the CLABSI rate per year (See Fig. 2). This CLABSI reduction rate is probably associated with INICC infection prevention interventions implemented during the last 24 years at these hospitals, which have been participating in this network of hospitals that voluntarily use the ISOS and the INICC multidimensional approach.Reference Higuera, Rosenthal, Duarte, Ruiz, Franco and Safdar23–Reference Alvarez-Moreno, Valderrama-Beltran and Rosenthal29 On the other hand, CLABSI rates increased twice: once in 2014 due to the addition of several new hospitals to the INICC network that had significantly higher CLABSI rates, and again in 2020 due to the COVID-19 pandemic (See Fig. 2).
Fig. 2. CLABSI rate per 1,000 CL-days, per year.
The LOS was linked to a 3% daily increase in the CLABSI risk. Jeon et alReference Jeon, Neidell, Jia, Sinisi and Larson30 conducted a study to examine the role played by LOS as a CLABSI risk factor. They conducted logistic regression and observed a nonlinear increase in the hazard of BSI with increasing LOS. The association between a longer LOS and an increased risk of CLABSI can largely be explained by the increased LOS among those who have underlying morbidity and require invasive procedures.Reference Jeon, Neidell, Jia, Sinisi and Larson30
We detected an incremental risk of acquiring CLABSI of 4% per CL day. Rey et alReference Rey, Alvarez and De-La-Rua31 also found an association of CL days as a CLABSI risk factor.Reference Rey, Alvarez and De-La-Rua31
In this study, the CL type with the highest risk of CLABSI was internal jugular, followed by femoral. In contrast, Lorente et alReference Lorente, Henry, Martin, Jimenez and Mora15 found that the femoral site had higher risk for CLABSI than the jugular site. In another study, the femoral site is safer than the jugular site in patients with tracheostomy. These findings suggest that the use of tracheostomy in addition to the jugular site leads to a higher risk than the femoral site.Reference Lorente, Jimenez and Naranjo32
In this study, PICC was the CL with the lowest risk of CLABSI. Chopra et alReference Chopra, O’Horo, Rogers, Maki and Safdar33 conducted a meta-analysis analyzing the risk of CLABSI associated with PICC compared with central venous catheters (CVCs). In their study, PICCs were associated with a lower risk of CLABSI than were CVCs (RR, 0.62; 95% CI, 0.40–0.94).Reference Chopra, O’Horo, Rogers, Maki and Safdar33 Also, Hon et alReference Hon, Bihari, Holt, Bersten and Kulkarni34 conducted a meta-analysis analyzing rate of CLABSI between tunneled CVCs versus PICCs in adult home parenteral nutrition. In their study, PICC use was associated with a significantly lower rate of CLABSI (RR, 0.40; 95% CI, 0.19–0.83).Reference Hon, Bihari, Holt, Bersten and Kulkarni34
In our study, patients admitted to adult oncology and pediatric oncology ICUs had the highest risk of CLABSI. The CL–DU ratio, as a marker of severity of illness of patients, was highest at those types of ICUs,Reference Dudeck, Edwards and Allen-Bridson18 which could explain why these ICUs were associated with the highest risk of CLABSI.
We also noted that publicly owned facilities and teaching hospitals had a significantly higher risk of CLABSI than for-profit privately owned facilities. This finding is consistent with a previous study conducted in NICUs, in which the CLABSI rate per 1,000 CL days at university hospitals was 14.3 (95% CI, 12.9–15.7), the CLABSI rate at publicly owned facilities was 14.6; 95% CI, 11.0–19.1, and the CLABSI rate at for-profit, privately owned facilities was 10.8 (95% CI, 8.5–13.5).Reference Rosenthal, Lynch and Jarvis35
Additionally, middle-income countries had a significantly higher risk of CLABSI than high-income countries. This finding could be explained by the likelihood of lower-quality programs in LMICs compared with high-income countries.Reference Rosenthal3 In a previous study conducted in NICUs, analyzing the impact of the income level of the country and CLABSI, the CLABSI rate per 1,000 CL days in low-income countries was 37.0 (95% CI, 16.0–71.8) and the CLABSI rate in upper–middle-income countries was 17.6 (95% CI, 15.3–20.2).Reference Rosenthal, Lynch and Jarvis35 In another study conducted in PICUs, the CLABSI rate in LMICs was 12.4 (95% CI, 10.5–14.3) and the CLABSI rate in upper–middle-income countries t was 7.0 (95% CI, 6.3–7.9).Reference Rosenthal, Jarvis and Jamulitrat36 In both studies, the higher the income level of the country, the lower the CLABSI rate.
We did not detect an association between sex and CLABSI. This finding is consistent with other studies that also did not detect such association.Reference Dahan, O’Donnell and Hebert37
We did not detect an association between age and CLABSI, which is inconsistent with the study of Hsu et al,Reference Hsu, Chang and Tsai38 who identified age >65 years as a CLABSI risk factor. We most likely did not find such an association because we controlled for 11 independent variables that were more significantly associated with CLABSI risk than age.
Some of the CLABSI risk factors identified in our study are unlikely to change, such as the income level of the country, facility ownership, hospitalization type, and ICU type. However, some of the risk factors for CLABSI we identified can be modified: CL days, LOS, use of tracheostomy and use of internal jugular or femoral lines.
Based on our findings, we should focus on strategies to reduce CL use, reduce LOS, prefer PICC instead of internal jugular or femoral insertion, and implement an evidence-based set of CLABSI prevention recommendations, such as those recently published by the Society for Healthcare Epidemiology of America (SHEA), Association for Professionals in Infection Control and Epidemiology (APIC), and the Infectious Diseases Society of America (IDSA).Reference Buetti, Marschall and Drees13 Also, the very high rate of CLABSI prevalent in LMICsReference Rosenthal, Maki and Salomao1–Reference Rosenthal, Myatra and Divatia4 can be reduced by utilizing a strategy of monitoring compliance with recommendations and providing performance feedback to healthcare personnel, as has been demonstrated in several LMICs.Reference Higuera, Rosenthal, Duarte, Ruiz, Franco and Safdar23–Reference Alvarez-Moreno, Valderrama-Beltran and Rosenthal29
Our study had several limitations. First, because this study was part of a surveillance system in which hospitals voluntarily participated for free; thus, these findings are not representative of all hospitals in LMICs. Second, changes in personal or professional practices may have influenced risk over time. Third, changes to CLABSI definitions made by the CDC that we adopted immediately may have influenced outcomes. Fourth, the unequal contribution of data by the participating hospitals may have affected these findings. Fifth, more clinical and epidemiological data as well as water quality could be useful to characterize the situation. Finally, the IPPs of the participating hospitals did not collect information on disease severity scores; instead, we used the CL-DU ratio to assess the severity of illness, and we adjusted the analysis to account for this independent variable.
Conclusion
The importance and relevance of this study lie in the fact that (1) it is a prospective cohort study using a standardized form with a denominators and numerators validation system, (2) always using the updated criteria and CLABSI definitions of the CDC NHSN, (3) with a duration longer than any study about risk factors for CLABSI never published before, (4) with the participation of a number of ICUs and countries, collecting data on patients, patient-days, and CLABSIs that were never collected before for an analysis of risk factors for CLABSI, and (5) with multiple logistic regression analysis adjusting for risk factors for 11 variables that have never been analyzed simultaneously before. Undoubtedly, this scientific work provides evidence about the independent variables that are risk factors for CLABSI and can serve as a guide for practice at the hospital level and also be incorporated into recommendations, whether regional, national, or international.
Acknowledgments
Financial support
No financial support was provided relevant to this article.
Conflicts of interest
All authors declare that they do not have any financial and personal relationships with other people or organizations that could inappropriately influence (bias) their work. All authors declare that do not have potential competing interests, such as employment, consultancies, stock ownership, honoraria, paid expert testimony, patent applications/registrations, and grants or other funding.
