Macrolides: a novel risk factor for carbapenemase-producing Enterobacterales in intensive care units

Introduction: Carbapenemase-producing Enterobacterales (CPE) have emerged as a substantial cause of morbi-mortality worldwide, with a prevalence of approximately 5% in areas with high endemicity. However, available data may not be representative of developing countries, such as Ecuador. In this study, the incidence of CPE in Ecuador and risk factors for infection/colonisation were evaluated. Methodology: A prospective cohort study was performed from February to April 2016 in seven intensive-care units of Guayaquil, Ecuador. Samples were processed according to the Centers for Disease Control and Prevention laboratory protocol and the CHROMagar mSuper CARBA agar method. Resistance to carbapenems was defined according to Clinical and Laboratory Standards Institute breakpoints. A modified carbapenemase inactivation method was used to identify carbapenamase production phenotypically with molecular confirmation by multiplex polymerase chain reaction. Results: In total, 640 patients were enrolled. The incidence of CPE was 36.4% (N = 233). A multivariate analysis indicated that several factors were associated with CPE acquisition, included a long intensive care unit stay (OR 1.05; 95% CI 1.03–1.08; p < 0.01), tracheostomy (OR 3.52; 95% CI 1.90–6.75; p < 0.01), hospitalisation 3 months prior to admission (OR 2.07; 95% CI 1.17–3.71; p < 0.01), vancomycin use (OR 3.31; 95% CI 2.02–5.18; p < 0.01), and macrolide use (OR 3.31; 95% CI 1.43–7.76; p < 0.01). Conclusions: Macrolide use was a risk factor for CPE acquisition. This association should be evaluated further, especially in developing countries.


Introduction
Carbapenemase-producing Enterobacterales (CPE) have emerged as a substantial cause of morbi-mortality worldwide [1], with a prevalence of approximately 5% in high endemicity areas. However, these statistics may not be representative of developing countries, such as Ecuador, where CPE has shown rapid dissemination since 2010 [2][3][4]. Thus, the aims of this study were to determine the incidence of CPE colonisation/infection and related risk factors in patients in intensive-care units (ICUs) in Guayaquil, Ecuador.

Study design
A prospective cohort study was performed between February and April 2016 in seven ICUs in Guayaquil, Ecuador. All admitted patients were screened with inguinal/perineal swabs weekly and upon admission if they were transferred from other institutions or had known risk factors for CPE. Patients were excluded if they were infected or colonised with CPE at enrolment [5].  [7]. Escherichia coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853, and Klebsiella pneumoniae ATCC BAA-1705 were used as quality control strains. A modified carbapenemase inactivation method was used to identify carbapenamase production. All CRE were subcultured on trypticase soy agar (Oxoid, Basingstoke, UK) for 24 hours at 35°C in air. DNA was extracted using the Wizard Genomic DNA Purification Kit (Promega, São Paulo, Brazil) following the manufacturer's recommendations. DNA quality was assessed by analysing the ratio of absorbance at 260 nm/280 nm. Carbapenemase production was confirmed by the detection of the blaKPC, bla OXA-48 , bla VIM , bla IMP , and bla NDM genes by a multiplex polymerase chain reaction, as described previously, using a thermal cycler (BioRad, Hercules, CA, USA). The reaction conditions were as follows: initial denaturation at 94°C for 10 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 52°C for 30 s, extension at 72°C for 1 min, and a final extension at 72°C for 10 min. The PCR products were resolved by 1% agarose gel electrophoresis at 120 V for 30 min. The gel was stained with Syber Safe (Invitrogen, Carlsbad, CA, USA). Primers used for the reactions have been described previously [8].

Statistical analysis
Data analyses were performed using R version 3.6.2 (Vienna, Austria: R Foundation for Statistical Computing, 2019; available at: www.R-project.org). Chi-squared tests or Fisher's exact tests and Student's t-tests or Mann-Whitney U tests were used to compare categorical or continuous variables, respectively. Multicollinearity, Pearson's correlation coefficients, and variance inflation factors for the logistic regression model were analysed. The Wald test was used with a significance level of α = 0.05 and 95% confidence intervals (CIs) for the OR were obtained.

Results
Out of 640 patients enrolled in the study, 233 (36.4%) had at least one sample positive for CPE.  Table 1.
Risk factors for infection or colonisation with CPE before current admission in patients who tested positive after enrolment included previous hospitalisation (RR 1 Table 2).

Discussion
Prospective studies aimed at identifying risk factors for CPE colonisation/infection in patients in developing countries are scarce. The incidence and prevalence of CPE varies among geographical regions. Previous studies have reported values ranging for 0.6% in Asia to 35.2% in Uganda. In this study, we observed an incidence of 36.4%, which is higher than those in other countries in South America, like Brazil [9,4]. The length of ICU stay and invasive device use contribute to CPE infection [3,10], consistent with the results of this study. Tracheostomy was the only invasive device associated with CPE colonisation/infection in our multivariate analysis, which is inconsistent with the results of some previous studies [11,12]. Therefore, additional studies of ventilator-associated pneumonia as an important end point are needed. Carbapenems, fluoroquinolones, antipseudomonal penicillin, and broad-spectrum cephalosporins increase the risk of KPC colonisation [3]. We were unable to confirm that the administration of these drugs is associated with CPE based on a multivariate analysis. Fluoroquinolones and broad-spectrum cephalosporins are not frequently used in the hospitals studied; thus, a limited number of patients received these antimicrobials, making statistical analyses difficult. We obtained two unexpected results: i) vancomycin administration increased CPE-colonisation two-fold and ii) the presence of macrolides was a risk factor. Vancomycin induces drastic and consistent changes in the human intestinal microbiota, with a decrease in gram-positive bacteria (mainly Firmicutes) and a compensatory increase in gram-negative bacteria. Importantly, several genera belonging to the phylum Proteobacteria (i.e., Escherichia/Shigella and Klebsiella) increase after vancomycin administration [13,14]. This effect could explain the detection of CPE in ICU patients exposed to a short course of vancomycin treatment. The identification of macrolides as risk factors can be explained by the fact that azithromycin treatment can increase the selection of macrolide resistance genes in the gut microbiome [15], which acts as an important reservoir of various resistance genes. Different mechanisms underlying macrolide resistance have been identified in Enterobacterales, some of which can be transferred amongst microorganisms via mobile genetic elements. Transferable mechanisms of macrolide resistance (TMMR) are present in genetic structures that also include β-lactam resistance; for example, the IncH1 plasmid from C. freundii, which might also be involved in macrolide extrusion, can harbour the blaNDM-1 gene [16]. Detailed studies of these resistance mechanisms are needed to better understand the relationship between macrolides and CPE.
The discovery of macrolide use as a risk factor has critical implications, since clarithromycin is a first-line treatment in severe community-acquired pneumonia and vancomycin is included as part of the empiric treatment in septic shock and high-risk febrile neutropenia [17]. Another macrolide, azithromycin, was initially considered for the treatment of COVID-19-related pneumonia and the collateral effects of this drug should be carefully addressed during the pandemic [18]. Our results confirm a report by the World Health Organization in 2017, suggesting that the use of these drugs should be prioritised for monitoring as part of an appropriate antimicrobial stewardship program [17]; these programs are scarce in developing countries, such as Ecuador. Thus, modifying the recommended empirical antimicrobial treatments based on the infectious diseases processes of each patient and improving our understanding of the clinical and molecular epidemiological properties as well as resistance mechanisms are key goals. Patients colonised or infected with CPE in our study had a significantly higher mortality rate than that of controls. This is an important finding, since most studies have focused on infection and its relationship with higher mortality rates. In one study, colonisation did not increase ICU mortality, whereas KPC infection did [11]. Our findings require further research for validation and highlight the importance of early identification and isolation [19]. The difficulty in obtaining reliable information on antimicrobial exposure and the use of invasive procedures outside of the study centres was a limitation of our study. We did not collect information about other microorganisms and related infections, which could result in selection bias. Finally, the short study period and the composite of variables related to colonisation and infection could have affected the identification of factors that determine the shift from colonisation to infection.

Conclusions
Although risk factors associated with CPE have been studied extensively in developed countries, it is imperative to evaluate these factors in developing countries with different public health contexts. Our study contributes to the few reports indicating that tracheostomy use as well as vancomycin and macrolide administration are risk factors for CPE, at least in developing countries. This information highlights the importance of including macrolide and vancomycin as WATCH antibiotics in antimicrobial stewardship programs and the necessity of reviewing empirical treatment recommendations. In this sense, there is a need for more research to improve antimicrobial stewardship programs and ultimately to prevent the spread of CPE.