Breakthrough bacteraemia due to tigecycline-resistant Escherichia coli with New Delhi metallo- -lactamase (NDM)-1 successfully treated with colistin in a patient with calciphylaxis

Sir, We present the case of a patient with calciphylaxis and co-infection with New Delhi metallo-β-lactamase (NDM)-1-producing Escherichia coli and Klebsiella pneumoniae, both susceptible to tigecycline and colistin. While receiving tigecycline, the patient developed a bacteraemia due to the NDM-1-positive E. coli, now showing tigecycline resistance. The infection was successfully treated with intravenous colistin. NDMs confer resistance to carbapenems. The NDM-1 enzyme was first recognized in 2008 in K. pneumoniae and E. coli isolates from a patient in Sweden who had been hospitalized in New Delhi.1 Bacteria with NDM enzymes have since been reported worldwide, often related to travel or hospitalization in the Indian subcontinent.2 NDM-1 represents an emerging therapeutic challenge; nevertheless, there are few descriptions of treatment experience. Calciphylaxis is a rare disorder, characterized by calcification of arterioles, leading to tissue ischaemia and necrosis.3 A 59-year-old patient with a history of type II diabetes mellitus presented to our hospital with a 2 month history of bilateral thigh ulcerations. The patient had visited Kenya and India in the year prior to admission; however, there was no reported contact with healthcare facilities during either visit, both for the patient and for close contacts. On admission, the lesions were clinically infected. Empirical treatment was with intravenous (iv) flucloxacillin and benzylpenicillin, then piperacillin/tazobactam, without clinical response. Biopsy of the lesions led to a histological diagnosis of calciphylaxis. A biopsy sample was cultured on standard media, with E. coli, K. pneumoniae, methicillin-resistant Staphylococcus aureus (MRSA) and Pseudomonas aeruginosa identified using the BD Phoenix automated identification system (Oxford, UK). The E. coli and K. pneumoniae were resistant to meropenem (confirmed by Etest) and ertapenem, and to penicillins, oxyimino-cephalosporins and aminoglycosides. They remained susceptible to tigecycline and colistin. Accordingly, antimicrobial therapy was altered to 50 mg of tigecycline twice daily iv (after a 100 mg loading dose) and 500 mg of ciprofloxacin twice daily orally to cover the P. aeruginosa. Screening cultures from stool were negative for carbapenem-resistant Enterobacteriaceae. The K. pneumoniae and E. coli isolates were referred to the Antibiotic Resistance Monitoring & Reference Laboratory of the HPA for MIC determination by BSAC agar dilution and molecular investigation. Both were multiresistant, with susceptibility confirmed only to tigecycline and colistin (MICs both ≤0.5 mg/L); PCR and sequencing confirmed the blaNDM-1 gene in both isolates, carried by plasmids belonging to the A/C rep type, which gave closely similar profiles after digestion with the SacI enzyme. The 16S rRNA methylase gene rmtC was also detected in both isolates by PCR, conferring resistance to all clinically used aminoglycosides. The patient required prolonged hospitalization, interrupted only by 4 weeks at home. Four months after initial presentation, while being treated for ongoing infection of calciphylactic lesions, the patient developed fever. Blood cultures were taken and grew NDM-1-positive E. coli identical by PFGE to that isolated from the original lesions. Susceptibility to colistin remained, but the isolate was now resistant to tigecycline (MIC 8 mg/L; Table 1). At the time of this episode the patient was on 50 mg of tigecycline twice daily iv, and had by that point received 53 days of tigecycline, indicating development of resistance during prolonged tigecycline therapy. MICs in mg/L for NDM-1-producing E. coli isolated from the patient; the second was isolated from blood 4 months after the original was isolated from a calciphylactic skin lesion MICs in mg/L for NDM-1-producing E. coli isolated from the patient; the second was isolated from blood 4 months after the original was isolated from a calciphylactic skin lesion The patient was then commenced on colistin [4.5 million units (MU) twice daily iv]. After 10 days of this treatment, the dose was reduced to 3 MU twice daily in the light of rising serum creatinine and elevated colistin levels (trough 8.5 mg/L, peak 10.4 mg/L). On day 13 of colistin treatment, admission to the intensive care unit (ICU) was required for supportive treatment of sepsis. Meropenem and teicoplanin were added empirically to cover the previously isolated P. aeruginosa and MRSA. On day 15 of colistin therapy the patient developed generalized seizures. CT brain scan and CSF were normal. Colistin therapy was discontinued and the seizures stopped shortly afterwards, suggesting colistin-associated neurotoxicity, a recognized side effect.4 The patient returned to a single room within a general ward after 10 days on the ICU. Subsequent blood cultures were negative. Six months after initial presentation, the patient remains clinically stable. Despite prolonged admission, contact isolation prevented onward transmission of the NDM-positive bacteria within the hospital. This is the first carbapenemase-producing E. coli confirmed to be resistant to tigecycline by the national reference laboratory. An 8-fold reduction in the MIC of tigecycline for the resistant isolate when tested in the presence of the efflux inhibitor phenyl-arginine-β-naphthylamide (PAβN; at 40 mg/L) was observed (compared with only a 2-fold reduction for the susceptible isolate), which suggests up-regulated efflux as the resistance mechanism. Future studies will ascertain the specific pump(s) involved. The isolates remained susceptible to colistin, as do most with NDM enzymes.5,6 This case suggests that invasive infection with NDM producers can be successfully treated with colistin, albeit with the risk of significant toxicity, and further demonstrates the emerging threat of the NDM-type metallo-β-lactamases. We received the patient's verbal and written consent for their case to be published in a scientific journal for the purposes of medical education and research. No specific funding has been received. Data have been generated as part of the routine work of the Royal Free Hampstead NHS Trust and the Antibiotic Resistance Monitoring & Reference Laboratory of the HPA. N. W. has received research grants and conference support from most major pharmaceutical companies; none represents a conflict of interest with the content of this article. D. M. L. has received grants, speaking invitations and conference invitations from most major pharmaceutical companies and holds shares in GSK, Pfizer, AstraZeneca, Merck, Dechra and Eco Animal Health within diversified portfolios. S. M. has received conference support from numerous pharmaceutical companies. J. H. has received a travel grant funded by Wyeth pharmaceuticals, ESCMID and ISC to attend a conference. All other others—nothing to declare.