Geographical and temporal clustering of OXA-48-producing Escherichia coli ST410 causing community-onset urinary tract infection in Christchurch, New Zealand

Sir, Carbapenemase-producing Enterobacteriaceae (CPE) are a threat to global health and feature prominently in the WHO action plan on antimicrobial resistance.1 OXA-48-producing Escherichia coli are rare in New Zealand (NZ) with only eight isolates submitted to the national reference laboratory [Institute of Environmental Science and Research (ESR)] from 2009 to 2017.2 In 2017 in Christchurch, NZ, we recovered two unusual resistant E. coli with the same antibiogram from two patients with community-onset urinary tract infection within a 6 week period. Notably, the two patients lived in the same district of Christchurch. The isolates were identified as E. coli by MALDI-TOF (Bruker Daltonics, Melbourne, Australia) and susceptibilities were performed with Phoenix™ (BD Diagnostics, USA) using EUCAST breakpoints. Both isolates flagged as possible CPE; however, they had low MICs to meropenem (0.5 mg/L) and imipenem (2 mg/L). They remained susceptible to trimethoprim/sulfamethoxazole, nitrofurantoin, ceftriaxone, fosfomycin and gentamicin. In-house Carba NP gave a positive result, suggesting carbapenemase production.3 Testing by the GeneXpert Carba-R PCR revealed blaOXA-48-like genes in both isolates so we set out to investigate whether they were related using WGS. Ethics approval was not needed as this was carried out as part of epidemiological investigation on routinely collected clinical isolates. Nucleic acid was extracted using the QIAamp UCP Pathogen Mini Kit (QIAGEN, Hilden, Germany) and DNA libraries were created using the Nextera XT DNA Preparation Kit (Illumina, Australia) as recommended by the manufacturer. WGS was performed on the MiSeq platform (Illumina) using MiSeq Reagent Kit v2 (300 cycle). Bioinformatic analysis included SPAdes genome assembly and contig upload to the CGE web portal (http://www.genomicepidemiology.org/) for MLST, PlasmidFinder and ResFinder analysis. Isolates were ST410 and harboured blaOXA-48, blaDHA-1, qnrB4, dfrA17 and sul1 resistance genes. Plasmid analysis indicated the following replicon types were present; IncFIA, IncFIB, IncFIC, IncFII and Col156. Further analysis could be performed to characterize the plasmids using long-read technology. SNP analysis was performed using Snippy (https://github.com/tseemann/snippy) and revealed only two SNP differences between the isolates. One patient had a history of travel to Cambodia a few months earlier but the other had not travelled overseas. We could find no evidence that the two patients were known to one another. Analysing NZ national data, we found that patients with OXA-48-producing E. coli were more likely to have a link with Cambodia (travel, hospitalization or Cambodian origin) than patients with E. coli carrying OXA-48-like carbapenemase genes [3/8 (37.5%) versus 0/17 (0%), respectively (14 OXA-181; 1 OXA-232; 1 OXA-244; and 1 OXA-566), P = 0.0243, Fisher’s exact test]. OXA-48 is a relatively weak carbapenemase, conferring only slightly elevated carbapenem MICs that may remain within the susceptible range. This can make laboratory detection challenging. Furthermore, the lack of ESBL resistance may further compound the challenge of detecting OXA-48-producing E. coli; although our isolates carried an acquired ampC gene (blaDHA-1) they remained susceptible to ceftriaxone (MIC = 1 mg/L) and aztreonam (MIC = 1 mg/L). Both isolates were susceptible to cefepime (MIC ≤0.5 mg/L) and showed an inducible AmpC on phenotypic testing with flattening of zones between imipenem and cefotaxime. NZ is an island nation of 4.5 million people where both healthcare and household transmission of CPE have already been reported in different regions.4,5 Although the prevalence of CPE is still low, it is concerning that 9% of patients with OXA-48-like-producing CPE from 2009 to 2017 have no history of travel and presumably acquired the organism in NZ (ESR data). Our current observation suggesting community acquisition reinforces the rationale for developing a national response plan for CPE as advocated in a recent viewpoint article and agreed to by the NZ Ministry of Health.5,6 NZ has a complex mixture of public and private providers of diagnostic microbiology services so laboratory protocols for detecting OXA-48-positive CPE are likely to have variable sensitivity. Standardizing laboratory detection protocols for suspected CPE is therefore one of the issues that could be addressed through the development of a national plan. ST410 E. coli is a widespread clone, previously isolated from humans as well as livestock; meat (turkey meat, suckling pig and chicken giblets) in Germany;7 raw turkey from Spain;8 and raw milk from cows with mastitis in Southern Taiwan.9 Data on the prevalence of CPE in Cambodia are scarce; however, a study of bloodstream isolates between 2007 and 2010 in Phnom Penh found a high prevalence of ESBL producers, but no CPE.10 In conclusion, we report two temporally and geographically linked patients with community-onset urinary tract infection caused by the same OXA-48-producing E. coli clone. The association between OXA-48 and Cambodia in NZ suggests the Cambodian patient may have acquired the organism during recent travel and indirectly transmitted to the second patient via unknown community transmission pathways. It is possible that OXA-48-positive CPE are under-recognized in NZ due to the challenges of laboratory detection. Screening with adequate selective media for CPE is highly recommended. The forthcoming national response plan to CPE in NZ has potential to help address some of these issues. This work has been presented in part as a poster (PO16) at Antimicrobials 2018, Brisbane, Australia. We acknowledge Helen Heffernan and Una Ren at ESR for national-level data. This study was conducted as part of our routine work. None to declare.