Estimating the association between antibiotic exposure and colonization with extended-spectrum β-lactamase-producing Gram-negative bacteria using machine learning methods: a multicentre, prospective cohort study

Abstract

Extended-spectrum b-lactamaseIncidence rate Machine learning a b s t r a c t Objectives: The aim of the study was to measure the impact of antibiotic exposure on the acquisition of colonization with extended-spectrum b-lactamase-producing Gram-negative bacteria (ESBL-GNB) accounting for individual-and group-level confounding using machine-learning methods. Methods: Patients hospitalized between September 2010 and June 2013 at six medical and six surgical wards in Italy, Serbia and Romania were screened for ESBL-GNB at hospital admission, discharge, antibiotic start, and after 3, 7, 15 and 30 days. Primary outcomes were the incidence rate and predictive factors of new ESBL-GNB colonization. Random forest algorithm was used to rank antibiotics according to the risk of selection of ESBL-GNB colonization in patients not colonized before starting antibiotics. Results: We screened 10 034 patients collecting 28 322 rectal swab samples. New ESBL-GNB colonization incidence with and without antibiotic treatment was 22/1000 and 9/1000 exposure-days, respectively. In the adjusted regression analyses, antibiotic exposure (hazard ratio (HR) 2.38; 95% CI 1.29e4.40), age 60 e69 years (HR 1.19; 95% CI 1.05e1.34), and spring season (HR 1.25; 95% CI 1.14e1.38) were independently associated with new colonization. Monotherapy ranked higher als combination therapy in promoting ESBL-GNB colonization. Among monotherapy, cephalosporins ranked first followed by tetracycline (second), macrolide (fourth) and cotrimoxazole (seventh). Overall the ranking of cephalosporins was lower when used in combination. Among combinations not including cephalosporins, quinolones plus carbapenems ranked highest (eighth). Among sequential therapies, quinolones ranked highest (tenth) when prescribed within 30 days of therapy with cephalosporins. Conclusions: Impact of antibiotics on selecting ESBL-GNB at intestinal level varies if used in monotherapy or combination and according to previous antibiotic exposure. These finding should be explored in future clinical trials on antibiotic stewardship interventions.

Impact of Antibiotic Usage on Extended-Spectrum ß-Lactamase Producing Escherichia Coli Prevalence: A Time-Series Analysis of Tertiary Hospital Implementing Stewardship Program

“…Other studies have shown a strong relationship between the usage of third-generation cephalosporin and expression of ESBL phenotype [3,15,16]. Tacconelli et al [3] reported that cephalosporin monotherapy ranked first among all antibiotics in promoting colonization of ESBL gram-negative bacteria in rectal swabs through analyzing the patient-level data with machine learning methods. However, in our study, this correlation was not evident.…”

Machine Learning and Multidrug-Resistant Gram-Negative Bacteria: An Interesting Combination for Current and Future Research

“…In this section, we describe a case example of a recent study exploiting he use of ML algorithms for depicting the risk of colonization [19] by extended-spectrum β-lactamases (ESBL)-producing Enterobacterales (ESBL-PE), a type of MDR-GNB. Indeed, although colonization by MDR-GNB is not always followed by MDR-GNB infection (i.e., colonized patients may not develop infection), the latter usually follows the former.…”

“…The study performed by Tacconelli and colleagues is a good example of combining classical statistical techniques and ML algorithms to provide useful complementary information [19]. First, the authors used flexible parametric survival models (adjusted for colonization pressure and ward of hospital stay) and identified previous antibiotic exposure as one of the independent predictors of ESBL-PE colonization (hazard ratio 2.38, 95% confidence intervals [CI] 1.29-4.40, p = 0.006).…”

“…Eventually, this approach provided useful additional information, by ranking previous monotherapy to be more at risk of promoting ESBL-PE colonization than previous combination therapy, and cephalosporins to be the most influential single antibiotic class. Of note, the authors initially trained different ML algorithms (linear support vector machine (SVM) with radial basis function kernel, neural network, random forest, and others) and ultimately selected the random forest algorithm on the basis of the best performance on the cross-validation set in terms of accuracy, overfitting, and permutation significance [19]. This further testified to the role of thoughtful human guidance of optimizing the application of ML techniques to specific risk prediction tasks.…”

Impact of antibiotic usage on extended-spectrum β-lactamase producing Escherichia coli prevalence

“…Various classes of antibiotics have been described as risk factors for the prevalence of ESBL-producing organisms, including cephalosporins, carbapenems, and trimethoprim/sulfamethoxazole 2 , 3 . These previous studies were based on the concept that using antimicrobial drug results in selective pressure toward the emergence of resistance 4 .…”