Factors associated with acquisition of carbapenem-resistant Enterobacteriaceae

ABSTRACT Objective: to identify possible risk factors for acquisition of Enterobacterial strains with a marker for resistance to carbapenems. Methods: exploratory case-control study performed in hospital settings. The study sample consisted of patients with biological specimens that tested positive for carbapenem-resistant Enterobacteriaceae (cases), with the disk diffusion test and Etest, and controls with biological samples testing negative for carbapenem-resistant Enterobacteriaceae. In all, 65 patients were included: 13 (20%) cases and 52 (80%) controls. Results: the microorganisms isolated were Serratia marcescens (6), Klebsiella pneumoniae (4), and Enterobacter cloacae (3). Univariate analysis revealed that length of hospitalization prior to sample collection (p=0.002) and having a surgical procedure (p=0.006) were statistically significant. In the multivariable logistic regression model, both were still significant, with odds ratios of 0.93 (p = 0.009; 95% CI: 0.89 to 0.98) for length of hospitalization prior to sample collection, and 9.28 (p = 0.05; 95% CI: 1.01 to 85.14) for having a surgical procedure. Conclusion: shorter hospitalization times and increased surveillance of patients undergoing surgery could play a decisive role in reducing the spread of carbapenem-resistant microorganisms in hospital settings.


Members of the Enterobacteriaceae family are
Gram-negative microorganisms found in nature, and isolated from biological material, that colonize the gastrointestinal tract of humans as part of the normal microbiota of this organ system, making it a potential reservoir for these pathogens. Carbapenemresistant Enterobacteriaceae (CRE) have emerged as an important cause of nosocomial infections around the world, and are characterized by rapid, progressive dissemination (1) . They are currently an important worldwide public-health problem, as infections due to CRE result in a high mortality rate, with limited therapeutic options (2)(3) .
Production of β-lactamase enzymes that can hydrolize carbapenems (carbapenemases) is one of the main mechanisms of resistance in Enterobacteriaceae.
According to the existing classification, carbapenemases belong to molecular class A (Klebsiella pneumoniae carbapenemase -KPC), B (metallobetalactamases, of which the primary ones are types VIM, IMP and NDM), and D (the most important being type OXA-48) (4) . The KPC is one of the most epidemiologically important types, because of its worldwide dissemination (5) .
Carbapenemases can be transferred between different strains of bacteria, usually by small circular DNA (deoxyribonucleic acid) molecules known as plasmids (4) , which can replicate independently from chromosomal DNA, and allow genetic material to be exchanged between different genera and species of Enterobacteriaceae (6) . This horizontal transfer of genes can involve multiple pathogens, and become widespread in a hospital setting.
The molecular epidemiology of carbapenemresistant bacteria has been extensively investigated.
However, most of the available information comes from studies that investigated specific bacteria (7)(8)(9)(10) or specific types of infection (11)(12) . The risk factors associated with the transmission of resistant pathogens cannot be fully understood when investigations are limited to specific bacteria, because plasmids with resistant traits can be transferred between bacteria of different species. Investigations of infection and colonization by CRE should therefore be more general, and not specify the genus of the bacteria or the patient's clinical condition (1,(13)(14)(15)(16) . Hence, there is a need for a study with a more comprehensive case definition, to provide a better understanding of the risk factors for infection by these microorganisms, so that effective prevention and control measures can be implemented.
The aim of the present exploratory casecontrol study was to identify possible risk factors for acquisition of Enterobacterial strains with a marker for carbapenem resistance.

Study design
This case-control study involved patients seen at one public and one not-for-profit hospital (a nongovernmental non-profit facility serving the public For each case, four randomly selected individuals with laboratory tests negative for CRE or any other organism, who were in the same unit at the same time as the case (± 20 days), composed the matched controls. Individuals whose records contained less than 50% of the information needed were excluded from the study. Controls whose records had insufficient information were replaced by other randomly selected controls. Approval for the study was granted by the

Microbiological procedures
Cultures sent to LACEN/ES were first tested biochemically to investigate bacterial metabolism (Himedia, Mumbai, India) in order to identify the genus/ species of the bacteria isolated. The biochemical tests included glucose, sucrose and lactose fermentation; CO 2 production; motility; indole production; urea hydrolysis; lysine, arginine and ornithine decarboxylase activity; citrate and malonate utilization; phenylalanine deaminase activity; and H 2 S production (17) .
After PCR amplification, the products were

Variables
The following items were investigated as possible risk factors: gender; age; hospitalization during the previous 90 days; hospitalization in an intensive care unit; use of a catheter or other invasive device; surgery during the current hospitalization; underlying comorbidities; and antimicrobial agents used during the current hospitalization. Unless otherwise stated, the events and periods considered in the analysis occurred before the biological samples were collected.

Sampling
To increase the power of the study, four controls hospitalized in the same unit during the same period as the cases were randomly selected and assigned to each case, giving a total of 13 cases and 52 controls enrolled in the study. was assessed by the Hosmer-Lemeshow test (21) . The data were analyzed in Statistical Package for the Social Sciences (SPSS), version 17. Figure 1 shows the Enterobacteriaceae species isolated in the 13 cases. Table 1 shows demographic and clinical characteristics of cases and controls.

Results
Only three of the records selected had to be replaced because of incomplete data. Univariate analysis of the variables, analyzed for their association with the outcome represented by colonization or infection by CRE, showed that length of hospitalization prior to sample collection (p=0.002) and having a surgical procedure (p=0.006) were statistically significant (Table 1). All the variables that had a p-value of less than 0.2 in the initial stage were included in the logistic regression model (Table 2).

Enterobacteriaceae Sample Type
Klebsiella pneumoniae Urine (two isolates)

Serratia marcescens
Wound secretion (two isolates)   This large interval indicates that the estimates from the logistic model are probably unstable, due to the small number of non-surgical cases (only two out of thirteen).

Serratia marcescens Blood
Of the 13 CRE isolates tested for resistance genes using PCR, nine (69.2%) were positive for the bla KPC gene: four isolates of K. pneumoniae, three of Enterobacter cloacae and two of Serratia marcescens. prior to sample collection is reported to vary from two to four weeks (22)(23) . This finding can be used to characterize CRE infection as a late-onset nosocomial complication.
Having surgery was a risk factor for acquiring CRE.
This finding is in agreement with a previous study that described surgery as being more common in patients with CRE infection, and corroborates the finding that medical procedures play a significant role in increased susceptibility of hospitalized patients to certain infections (24) .