Logo: to the web site of Uppsala University

Background

Antibiotic heteroresistance is a common bacterial phenotype characterised by the presence of small resistant subpopulations within a susceptible population. During antibiotic exposure, these resistant subpopulations can be enriched and potentially lead to treatment failure. In this study, we examined the prevalence, misclassification, and clinical effect of heteroresistance in Escherichia coli bloodstream infections for the clinically important antibiotics cefotaxime, gentamicin, and piperacillin–tazobactam.

Methods

We conducted a retrospective cohort analysis of patients (n=255) admitted to in-patient care and treated for E coli bloodstream infections within the Uppsala region in Sweden between Jan 1, 2014, and Dec 31, 2015. Patient inclusion criteria were admission to hospital on suspicion of infection, starting systemic antibiotics at the time of admission, positive blood cultures for the growth of E coli upon admission, and residency in the Uppsala health-care region at the time of admission. Exclusion criteria were growth of an additional pathogen than E coli in blood cultures taken at admission or previous inclusion of the patients in the study for another bloodstream infection. Antibiotic susceptibility of preserved blood culture isolates of E coli was assessed for cefotaxime, gentamicin, and piperacillin–tazobactam by disk diffusion and breakpoint crossing heteroresistance (BCHR) was identified using population analysis profiling. The clinical outcome parameters were obtained from patient records. The primary outcome variable was length of hospital stay due to the E coli bloodstream infection, defined as the time between admission and discharge from inpatient care as noted on the physician’s notes. Secondary outcomes were time to fever resolution, admission to intermediary care unit or intensive care unit during time in hospital, switching or adding another intravenous antibiotic treatment, re-admission to hospital within 30 days of original admission, recurrent E coli infection within 30 days of admission to hospital, and all-cause mortality within 90 days of admission.

Findings

A total of 255 participants with a corresponding E coli isolate (out of 500 screened for eligibility) met the inclusion criteria, with 135 female patients and 120 male patients. One (<1%) of 255 strains was BCHR for cefotaxime, 109 (43%) of 255 strains were BCHR for gentamicin, and 22 (9%) of 255 strains were BCHR for piperacillin–tazobactam. Clinical susceptibility testing misclassified 120 (96%) of 125 heteroresistant bacterial strains as susceptible. The BCHR phenotypes had no correlation to length of hospital stay due to the E coli bloodstream infection. However, patients with piperacillin–tazobactam BCHR strains who received piperacillin–tazobactam had 3·1 times higher odds for admittance to the intermediate care unit (95% CI 1·1–9·6, p=0·041) than the remainder of the cohort, excluding those treated with gentamicin. Similarly, those infected with gentamicin BCHR who received gentamicin showed higher odds for admittance to the intensive care unit (5·6 [1·1–42·0, p=0·043]) and mortality (7·1 [1·2–49·2, p=0·030]) than patients treated with gentamicin who were infected with non-gentamicin BCHR E coli.

Interpretation

In a cohort of patients with E coli bloodstream infections, heteroresistance is common and frequently misidentified in routine clinical testing. Several negative effects on patient outcomes are associated with heteroresistant strains.

Background

Heteroresistance (HR) is a significant type of antibiotic resistance observed for several bacterial species and antibiotic classes where a susceptible main population contains small subpopulations of resistant cells. Mathematical models, animal experiments and clinical studies associate HR with treatment failure. Currently used susceptibility tests do not detect heteroresistance reliably, which can result in misclassification of heteroresistant isolates as susceptible which might lead to treatment failure. Here we examined if whole genome sequence (WGS) data and machine learning (ML) can be used to detect bacterial HR.

Methods

We classified 467 Escherichia coli clinical isolates as HR or non-HR to the often used β-lactam/inhibitor combination piperacillin-tazobactam using pre-screening and Population Analysis Profiling tests. We sequenced the isolates, assembled the whole genomes and created a set of predictors based on current knowledge of HR mechanisms. Then we trained several machine learning models on 80% of this data set aiming to detect HR isolates. We compared performance of the best ML models on the remaining 20% of the data set with a baseline model based solely on the presence of β-lactamase genes. Furthermore, we sequenced the resistant sub-populations in order to analyse the genetic mechanisms underlying HR.

Findings

The best ML model achieved 100% sensitivity and 84.6% specificity, outperforming the baseline model. The strongest predictors of HR were the total number of β-lactamase genes, β-lactamase gene variants and presence of IS elements flanking them. Genetic analysis of HR strains confirmed that HR is caused by an increased copy number of resistance genes via gene amplification or plasmid copy number increase. This aligns with the ML model's findings, reinforcing the hypothesis that this mechanism underlies HR in Gram-negative bacteria.

Interpretation

We demonstrate that a combination of WGS and ML can identify HR in bacteria with perfect sensitivity and high specificity. This improved detection would allow for better-informed treatment decisions and potentially reduce the occurrence of treatment failures associated with HR.