Central-line–associated bloodstream infection (CLABSI) is a challenge worldwide. Although interventions have led CLABSI incidence density to decrease, many institutions maintain high rates in cancer patients, suggesting that infection may be derived from other sources beyond central venous cateters (CVCs). Reference Metzger, Rucker and Callaghan1
In 2013, the Centers for Disease Control and Prevention (CDC) developed a new definition called mucosal barrier injury laboratory-confirmed bloodstream infection (MBI-LCBI), which applies to patients with bacteremia and CVC (suspected CLABSI) but whose bacteremia has a probable gastrointestinal source, with association of mucosal barrier breakdown and enteric microorganism isolation. 2 Despite the clear importance of using the MBI-LCBI criteria, it is important to consider that this definition uses epidemiological criteria. A portion of these cases may actually be a CVC infection. Reference Vaughan, Ross and Gilman3
We evaluated the interference of the use of the MBI-LCBI criteria on the CLABSI incidence density in patients with solid tumors and hematologic malignancies, and we evaluated the proportion of cases of catheter-related bloodstream infection (CRBSI) among those classified as MBI.
Methods
Retrospective cohort study conducted at the São Paulo State Cancer Institute (ICESP), including cases from January 2015 to December 2017. ICESP is a 500-bed adult oncology university hospital, with 103 doctors’ offices and 18 operating rooms. Most beds are for patients with solid tumors. It also has an emergency room and a day hospital and provides chemotherapy and radiotherapy services. It does not provide hematopoietic stem cell transplantation (HSCT) service.
Data of CLABSI cases diagnosed were reviewed, and after reclassification considering the MBI-LCBI criteria, and finally these defined as MBI-LCBI were reclassified again using the CRBSI criteria.
For the diagnosis of CLABSI and MBI-LCBI, we used the National Healthcare Safety Network criteria of 2016. 2 The diagnosis of CRBSI followed the Infectious Diseases Society of America criteria: isolation at the tip of a catheter with >15 colony-forming units per plate and the growth of a microorganism of the same species and antibiogram profile in peripheral, or the growth of the same microorganism in a CVC at least 120 minutes faster than in the peripheral sample. Reference Mermel, Allon and Bouza4 We defined multidrug-resistant bacteria as vancomycin-resistant Enterococcus spp; methicillin-resistant Staphylococcus aureus; linezolid-resistant coagulase-negative Staphylococcus; carbapenem-resistant gram-negative bacilli (GNB), extended-spectrum-ß-lactamase (ESBL)-producing Enterobacteriaceae. 5 Neutropenia was defined by neutrophils ≤500/mm Reference Vaughan, Ross and Gilman3 or ≤1,000/mm Reference Vaughan, Ross and Gilman3 with expected decrease to ≤500 in the following 2 days. 6
We analyzed the following outcomes: (1) CLABSI infection density, without considering MBI and not excluding those classified as MBI; (2) CLABSI infection density, considering MBI and excluding those classified as MBI; and (3) MBI-LCBI infection density and CRBSI ratio among those classified as MBI-LCBIs. Each outcome was evaluated for the entire population and separately for patients with solid tumors and hematologic malignancies and for patients in wards and ICUs.
Results
We identified 339 cases of CLABSI in 322 patients, 232 (68.4%) in patients with solid tumors and 107 (31.5%) cases of CLABSI in patients with hematologic malignancies. Of the 339 CLABSIs, 53.1% occurred in the ICU, 17.4% occurred in the hematology ward, 16.2% occurred in clinical oncology, and 13.3% occurred in surgical oncology. The main diagnoses of base diseases were leukemia (16.5%), colorectal neoplasia (10.0%), bowel neoplasia (9.7%), and lymphoma (9.7%). Moreover, 450 microorganisms were isolated in blood cultures: 223 (48.9%) were gram-negative bacteria, 138 (30.6%) were gram-positive cocci (GPC), and 92 (20.5%) were fungi.
Applying the MBI-LCBI criterion, 51 episodes of MBI-LCBI (15.0%) were diagnosed in 50 patients. Among them, 6 cases (11.8%) were in patients with solid tumors and 45 (88.2%) occurred in patients with hematologic malignancies. Furthermore, 17 (33.3%) occurred in the ICU and 34 (66.7%) in the hematology ward. The oncology diagnoses of patients with MBI-LCBI included 35 leukemias, 9 lymphomas, 2 anorectal neoplasms, 1 bowel neoplasms, 1 neoplasm of unknown primary site, 1 male reproductive tract neoplasia, 1 lung neoplasia, and 1 multiple myeloma.
In the 51 MBI-LCBIs, 62 microorganisms were isolated; 40 (64.5%) were GNB, 15 (24.2%) were GPC, and 7 (11.3%) were fungi. Also, 31 (50.0%) were multidrug-resistant isolates and 27 (87.1%) occurred in patients with hematologic malignancies (Table 1).
Table 1. Distribution of Isolated Microorganisms in Blood Cultures of the 51 MBI-LCBI and Multidrug Resistance
Note. MBI-LCBI, mucosal barrier injury laboratory-confirmed bloodstream infection; MR, multidrug resistant.
Of the 51 MBI-LCBI, 10 (19.6%) could be classified as CRBSI when the specific criterion was applied. All cases occurred in patients with hematologic malignancies. Moreover, 11 microorganisms were isolated: 6 Klebsiella pneumoniae, 3 Escherichia coli, 1 Enterococcus spp and 1 Streptococcus salivarius. Also, 5 isolated microorganisms (45.5%) were multidrug resistant: 1 Enterococcus spp and 4 K. pneumoniae. The calculated incidence densities are shown in Table 2.
Table 2. Incidence Density of CLABSI and MBI-LCBI and Proportion of CRBSI: Patients with Solid Tumors or Hematology Malignancies
Note. CLABSI, central-line–associated bloodstream infection; MBI-LCBI, mucosal barrier injury laboratory-confirmed bloodstream infection; CRBSI, catheter-related bloodstream infection; ICU, intensive care unit; W, ward.
Discussion
In this study, 15% of CLABSIs were classified as MBI-LCBI when we applied the criteria. Few published studies have reviewed CLABSIs while applying the MBI-LCBI criteria, but some studies have shown that 71% of the CLABSIs can be classified as MBI-LCBIs. Reference Metzger, Rucker and Callaghan1
Overall, 88% of MBI-LCBIs occurred among inpatients with hematologic malignancies. This finding is similar to that of Islas-Muñoz et al, Reference Islas-Muñoz, Volkow-Fernández and Ibanes-Gutiérrez7 who reported that MBI-LCBI is most frequent in patients with hematologic malignancies.
When MBI-LCBI was removed, the CLABSI infection density decreased by 15.02% below the literature data, which reported reductions of 38%–46% in adult cancer patients. Reference Epstein, See, Edwards, Magill and Thompson8,Reference See, Soe, Epstein, Edwards, Magil and Thompson9 However, when we stratified the population, we observed reductions of 2.62% in the population with solid tumors and 42.03% in that with hematologic malignancies. Assessing only the cases in the ICU, the CLABSI incidence density decreased by 9.45% when MBI-LCBI was removed; the reduction was 4.85% in patients with solid tumors and 22.9% in those with hematologic malignancies. See et al Reference See, Soe, Epstein, Edwards, Magil and Thompson9 found a 30% reduction in CLABSI incidence density among adult ICU cancer patients.
Quantification of the impact of MBI-LCBI removal from CLABSI data is needed to allow an accurate interpretation of the trends and to more accurately target preventive measures, in addition to being aware of the CVC incidence.
Despite the interest in applying the MBI criteria, caution should be exercised not to completely remove CVC of focus of attention. 2 When using the MBI-LCBI criterion, which is essentially epidemiological, we wondered whether CVC was not the focus of infection. Kamboj et al Reference Kamboj, Blair, Bell, Sun, Eagan and Sepkowitz10 reviewed 32 episodes of CLABSI in HSCT patients, and 50% were MBI-LCBIs. Of the cases analyzed, 82% were caused by bacterial translocation, but 18% were actually CRBSIs, that is, the CVC was the focus of the infection. In our study, 19.6% of MBI-LCBI met criteria for CRBSI, demonstrating the chance of error when using the MBI-LCBI criterion because they are real CVC-focused infections.
Our study had several limitations. First, it was retrospective, with the possibility of observational bias. Additionally, the institution has a higher proportion of beds for patients with solid tumors than most cancer hospitals, which may limit the extrapolation of these results. Due to hospital characteristics, patients undergoing HSCT were not included, restricting cases defined as MBI-LCBI to the criterion of neutropenia.
On the other hand, to the best of our knowledge, this is the first Brazilian study to investigate MBI-LCBI by comparing patients with solid tumors and hematologic malignancies and using the CRBSI criteria to confirm the relationship of infection with the invasive device.
In conclusion, our results demonstrate that, in patients with hematologic malignancies, the use of the MBI-LCBI criteria significantly reduced the CLABSI incidence density, both in wards and in the ICU. We detected no impact on individuals with solid tumors. On the other hand, a significant proportion of cases classified as MBI-LCBI were actually CRBSI. The use of the combined criteria could measure the risks associated with greater accuracy and therefore better guide prevention and control measures.
Acknowledgments
We thank to all the professionals who work at the Hospital Infection Control and Infectious Diseases Service of Instituto do Câncer do Estado de São Paulo
Financial support
No financial support was provided relevant to this article.
Conflicts of interest
All authors report no conflicts of interest relevant to this article.
