Infectious Disease · Medical Literature
[Articles] Gene transfer from NDM-5-producing and OXA-48-producing Enterobacter hormaechei ST79 on contaminated dicloxacillin capsules to other Enterobacterales in Europe, 2020–23: a retrospective, observational, molecular epidemiological study
Abstract
This study describes how a clonal CPE outbreak caused by a contaminated medicine evolved into a complex plasmid-mediated outbreak involving multiple Enterobacterales species. Most patients related to the outbreak did not present with the original E hormaechei ST79 outbreak strain and were therefore not identified using standard outbreak detection methods. These findings highlight the importance of using plasmid-focused approaches in outbreak investigations.
Clinical relevance
Did the clonal outbreak of NDM-5/OXA-48-producing Enterobacter hormaechei ST79 linked to contaminated dicloxacillin capsules in Europe (2020–2023) evolve beyond clonal spread into a plasmid-mediated, multispecies outbreak that would be missed by conventional clone-focused surveillance methods? __SNAPSHOT__ {"studyDesign":"Retrospective, observational, molecular epidemiological study using nationwide WGS-based surveillance data","setting":"Denmark (nationwide); all five regional Departments of Clinical Microbiology reporting to the National Reference Laboratory for Antimicrobial Resistance, Statens Serum Institut, Copenhagen. Population coverage approximately 6 million residents.","sampleSize":"All confirmed CPE isolates from the Danish national surveillance submitted between Jan 1, 2014, and Oct 1, 2023; 160 isolates selected for long-read sequencing based on in-silico screening; total outbreak-linked isolates: 53 (across 39 patients). Exact denominator of the full surveillance database not reported in the abstract.","population":"Patients with confirmed carbapenemase-producing Enterobacterales (CPE) identified through mandatory national sequence-based surveillance; one isolate per species per patient per 12 months included. Outbreak patients were significantly older and predominantly female compared with non-outbreak control patients.","intervention":"Plasmid-focused genomic outbreak investigation: in-silico k-mer screening of the national CPE database for unique fragments of two outbreak-associated plasmids (pDcap_OXA-48 [71,966 bp IncL] and pDcap_NDM-5 [45,048 bp IncX3]), followed by long-read sequencing of 160 candidate isolates to obtain complete plasmid sequences, supplemented by qualitative conjugation transfer experiments.","comparator":"Conventional clonal outbreak detection methods (cgMLST-based strain matching, as used in the initial outbreak investigation); also compared against a non-outbreak CPE patient control group for epidemiological characteristics.","primaryOutcome":"Identification of all patients linked to the outbreak via either the original E. hormaechei ST79 clonal strain or either/both of the two outbreak-associated plasmids (pDcap_OXA-48 and pDcap_NDM-5) in other Enterobacterales species; characterisation of interspecies plasmid transfer events.","followUp":"Surveillance data spanning July 1, 2020 (earliest linked patient) to Oct 1, 2023; contaminated product distributed approximately July 1, 2020 to Feb 6, 2023.","fundingSource":"Not reported in the available abstract/full-text excerpt"}
Key findings
Initial clonal outbreak investigation (conventional cgMLST-based methods) identified 11 patients carrying the E. hormaechei ST79 outbreak strain; the Lancet Microbe study expanded this to 15 patients carrying the outbreak strain (including 4 newly identified). Using plasmid-focused screening, 53 total outbreak-linked isolates were identified across 39 patients — of whom 24 additional patients presented with other Enterobacterales species carrying one or both outbreak plasmids (pDcap_OXA-48 and/or pDcap_NDM-5) but not the original E. hormaechei ST79 strain. Conventional methods detected only 16 of the 53 outbreak isolates (30.2%, calculated); plasmid-specific screening identified the remaining 37 (69.8%, calculated). The pDcap_OXA-48 (IncL) plasmid transferred to seven distinct Enterobacterales species; pDcap_NDM-5 (IncX3) transferred to four species. Both plasmids were confirmed to be self-transmissible by qualitative conjugation experiments. Interspecies plasmid transfer was presumed to occur within individual patients, as only the E. hormaechei ST79 strain — not other Enterobacterales — was isolated from the implicated capsules. In four cases, the outbreak-associated plasmids were structurally modified compared to the reference sequences, raising the possibility of additional missed cases. The contaminated medical product (dicloxacillin 500 mg capsules) had been administered to approximately 79,000 individuals in Denmark between July 1, 2020 and Feb 6, 2023. No mortality attributable specifically to the plasmid-transferred isolates is reported in the abstract. No p-values, confidence intervals, NNT/NNH, or formal statistical comparisons between clonal and plasmid-focused detection yields are reported in the available abstract.
Practice implications
This study established that a pharmaceutical-contamination–driven clonal CPE outbreak can evolve into a complex plasmid-mediated, multispecies outbreak that is largely invisible to standard clonal surveillance approaches: the majority of outbreak-linked patients (approximately 62%, calculated) would not have been identified using conventional cgMLST-based methods because they did not harbour the index strain. The findings demonstrate in vivo horizontal transfer of carbapenemase-carrying plasmids from the ingested contaminating organism to recipient gut Enterobacterales, spanning at least four to seven species. The study did not establish clinical outcomes (infection vs. colonisation rates, mortality, or treatment failure) specifically attributable to the plasmid-transferred carbapenemase genes in the newly identified patients. It did not determine what proportion of the ~79,000 exposed individuals developed clinical infection, colonisation, or asymptomatic carriage — asymptomatic patients were not screened. Findings apply primarily to immunocompromised, older, or comorbid patients receiving dicloxacillin who subsequently develop CPE-positive cultures; the degree to which these findings generalise to other contaminated pharmaceutical exposures or healthcare transmission settings is not established by this single-source outbreak. The absence of outcome data and the retrospective observational design prevent causal attribution of clinical harm to the plasmid transfer events. The critical practical implication is that surveillance systems relying solely on organism-level (clonal) detection will systematically undercount carbapenemase-related outbreaks when plasmid-mediated spread has occurred. Only the abstract and ScienceDirect excerpt were available for review; full data tables, multivariate analyses, and supplementary figures were not accessible. __LIMITATIONS__ Only CPE isolates from patients already detected through routine clinical sampling were included; asymptomatic colonised patients among the ~79,000 exposed individuals were not screened, leading to probable significant undercount of the true outbreak size.,Retrospective design using existing surveillance data; causal inference about in vivo plasmid transfer is inferential — direct microbiological proof of transfer within individual patients (e.g., sequential sampling before and after transfer) is not reported.,Plasmid-specific screening using short-read k-mer alignment may produce false negatives if plasmid backbones have diverged sufficiently; the authors acknowledge this risk explicitly, noting four cases with structurally modified plasmids.,Epidemiological interviews were only conducted for 24 of 39 identified patients, as Danish legislation precluded interviewing patients without a confirmed microbiological link; this introduces ascertainment and selection bias.,No clinical outcome data (infection vs. colonisation, mortality, attributable morbidity) are reported for the plasmid-outbreak patients, limiting assessment of clinical significance beyond microbiological findings.,Funding source not reported; potential conflicts of interest cannot be assessed.,Single-country (Denmark) setting with mandatory national CPE WGS surveillance — a resource not available in most health systems — limits direct generalisability of the plasmid-focused approach to settings lacking universal genomic surveillance infrastructure.,No formal statistical testing or effect-size estimates comparing detection yield of plasmid-focused vs. conventional approaches are reported in the available abstract.,Only the abstract and excerpts were available for full review; detailed methods, sensitivity/specificity of the k-mer screening algorithm, and complete data tables were not accessible. __BOTTOM_LINE_CATEGORY__ potentially_practice_informing __BOTTOM_LINE__ A contaminated pharmaceutical product seeded a dual-carbapenemase (NDM-5/OXA-48) E. hormaechei ST79 strain into thousands of patients; the carbapenemase-carrying plasmids subsequently transferred to up to seven other Enterobacterales species in vivo, expanding the outbreak to 39 patients — the majority of whom carried non-index organisms invisible to standard clonal surveillance. These findings provide compelling molecular epidemiological evidence that outbreak investigations involving carbapenemase-producing organisms should routinely incorporate plasmid-level genomic analysis in addition to cgMLST-based clonal typing. Adoption of plasmid-focused surveillance methods will require investment in long-read sequencing infrastructure and standardised analytic pipelines, but represents a potentially practice-changing advance in CPE outbreak detection.