Sir, Clonally diverse NDM-1-producing Escherichia coli (ST405, ST131, ST156 and ST101) have been reported globally, including on the Indian subcontinent. Indeed, there may be an association between isolates of NDM-1-producing E. coli ST101, phylogroup B1, that have been identified in Australia, Canada, Germany, the UK and Pakistan. In December 2011, NDM-1-producing E. coli (CREC-36) was identified from a patient who was hospitalized in a 260 bed tertiary care hospital in Korea. Medical records were reviewed and 114 stool or rectal swab samples were taken from 67 patients on Ward 1 (19 rooms), 11 patients on the intensive care unit (ICU) and 25 healthcare workers (doctors, nurses and caregivers) working on Ward 1. We screened patients in the ICU because the first patient had been treated in the ICU for 1 month before she was transferred to room 112. In addition, 23 environmental swabs were collected from washbasins, medicine trays and linen. All samples were cultured in 5 mL of tryptic soy broth containing 10 mg meropenem discs and processed using the CDC protocol. The modified Hodge test and the double-disc synergy test were performed on representative colonies grown on MacConkey agar. The blaNDM-1 gene was identified by PCR and sequencing using the primers NDM-1 F (5′-CAATATTATGC ACCCGGTCG-3′) and NDM-1 R (5′-ATCATGCTGGCCTTGGGGAA-3′). Three more patients who had shared a room (no. 112) with the first patient were also colonized by NDM-1-producing E. coli (CREC-45, CREC-47 and CREC-49). All four patients were elderly (.75 years) and had been in hospital for between 1 and 7 months before the NDM-1-producing E. coli were isolated. No NDM-1-producing E. coli were detected in samples from other patients, healthcare workers or the environment. NDM-1-positive individuals were isolated until decolonization had been confirmed by three consecutive negative stool cultures collected 1 week apart. None of the patients received treatment to eradicate the NDM-1 E. coli. Two patients died from their underlying disease. The remaining two were decolonized 21 and 53 days after the first isolation. No further cases of NDM-1 were detected. All the isolates were genetically related (PFGE identifying one pulsotype with .90% similarity), and the clone was identified by multilocus sequence typing (MLST; http://mlst.ucc.ie/mlst/dbs/ Ecoli) as ST101, phylogenetic group B1. E. coli ST101 is an international clone, frequently associated with the Indian subcontinent. However, the four patients had no epidemiological link to the Indian subcontinent as they had been hospitalized in Korea for a long time due to other underlying diseases. All patients shared an epidemiological link in terms of transmission. Although none of the healthcare workers harboured NDM-1, personal contact between the caregivers and the patients resident in the same room is the most likely transmission route; the NDM-1-colonized patients were bedridden, the beds were very close to each other and no NDM-1 cases were detected in any other rooms on Ward 1. Transmission was probably caused by poor standards of hygiene in the hospital; there were no sinks or alcohol-based hand cleansers in the room. Neither did the hospital monitor the hygiene practices of the healthcare workers. The MICs of antimicrobial agents for the NDM-1-producing isolates were determined using the agar dilution method and Etest. The susceptibility test results were interpreted according to CLSI or EUCAST (www.eucast.org; for colistin and tigecycline) guidelines. All NDM-1-producing isolates except CREC-49 showed high MICs of carbapenems (≥32 mg/L) and multidrug resistance; however, they were susceptible to tigecycline and colistin (Table 1). Using a blaNDM-1 probe, we identified the blaNDM-1 gene by Southern blotting of S1 nuclease-digested total DNA. The blaNDM-1 gene was identified in the 170, 180 and 120 kb plasmids. The discrepancy in blaNDM-1-bearing plasmid size may be because the plasmids are quite unstable and prone to rearrangement, as previously reported. Although the conjugal transfer of plasmids carried by the NDM-1-producing isolates failed, the blaNDM-1-harbouring plasmids from the four isolates were successfully transformed into electrocompetent E. coli HB101. PCR-based replicon typing of the major plasmid incompatibility groups showed that the plasmids belonged to the IncA/C incompatibility group. All transformants co-harboured blaCTX-M-15, blaCMY-4, blaDDHA-1 and armA. The genetic elements surrounding blaNDM-1 were examined by overlapping PCR mapping. The overall structure was ISAba125–blaNDM-1–bleMBL–trpF–blaDDHA-1–ampR–hypA– Dqac–sul1–ISCR1–tnpU–armA. The sequence spanning from ISAba125 to armA was identical to that in the E. coli DVR22 plasmid (IncHI1), and the sequence from blaNDM-1 to armA was identical to that in pNDM-HK (IncL/M). The sequence sul1–ISCR1–tnpU–armA was also identified in the IncA/C