Activity of temocillin against prevalent ESBL- and AmpC-producing Enterobacteriaceae from south-east England

Sir, Temocillin is the 6-α-methoxy derivative of ticarcillin. This modification increases stability to β-lactamases, including AmpC and extended-spectrum types, but reduces activity against Pseudomonas aeruginosa and removes affinity for the essential penicillin-binding proteins of Gram-positive bacteria and Bacteroides spp., giving a narrow-spectrum agent with very consistent activity against Enterobacteriaceae.1 Temocillin was marketed by Beecham Pharmaceuticals in the UK in the 1980s but achieved little commercial success and was withdrawn, though it remained available via the manufacturer's medical department. It continued to be marketed in Belgium, where it has been widely used in microbiologically directed therapy against infections known to be caused by pathogens with potent β-lactamases. Presently licensed to Eumedica, temocillin is being re-launched in the UK, and we assessed its activity against current ESBL-producing and AmpC-hyperproducing Enterobacteriaceae. The isolates used were from a recent survey of cephalosporin resistance in London and south-east England.2 They were consecutively collected Enterobacteriaceae found resistant at hospital Trust laboratories to one or more of cefpodoxime, cefotaxime and ceftazidime. MICs were determined centrally using the British Society for Antimicrobial Chemotherapy (BSAC) method3 and isolates were further characterized by a combination of interpretative reading4 and multiplex PCRs for blaCTX-M and acquired blaAmpC genes,5,6 and with a single PCR to identify members of Escherichia coli strain A, a particularly prevalent clone with blaCTX-M-15 linked to an IS26 element.7 In total, 846 survey isolates were tested with temocillin (see Table 1). These comprised (i) isolates with ESBL phenotypes, positive or not for blaCTX-M; (ii) isolates with AmpC phenotypes, with chromosomal or acquired (mostly blaCIT in E. coli) types and (iii) Klebsiella oxytoca inferred to hyperproduce K1 enzyme on the basis of high-level resistance to cefuroxime and aztreonam but retained susceptibility to ceftazidime. We cannot exclude the possibility that some isolates had additional mechanisms beyond those inferred or found. In particular, we are aware that some of the AmpC-derepressed enterobacters have blaSHV-12, although little or no cefepime/clavulanate synergy was apparent (M. J. Ellington, Antibiotic Resistance Monitoring and Reference Laboratory, Centre for Infections, personal communication). Apart from identifying members of E. coli strain A, we did not investigate the clonality of the collection; nevertheless, they were from 16 sites, with a cap of 100 isolates per site, limiting the distortions that could arise from multiple inclusion of outbreak strains. We also tested 13 non-survey Klebsiella pneumoniae isolates with resistance to ertapenem (MICs ≥ 8 mg/L) and reduced susceptibility or resistance to meropenem (MICs 4–16 mg/L; less so to imipenem: MICs 1–8 mg/L) arising via combinations of CTX-M ESBLs and loss of porins and/or up-regulated efflux. We are receiving small but increasing numbers of such isolates via the national reference service, although they remain too rare to register in surveillance studies. MIC distributions of temocillin for β-lactamase-producing Enterobacteriaceaea Modal values are indicated in boldface. In addition to the species groups detailed, this total also includes two Hafnia spp. and one Pantoea sp. with hyperproduced AmpC enzymes and one Proteus mirabilis with an ESBL. Includes 13 isolates with acquired CIT types. Includes 79 isolates representative of strain A, also detailed separately below. Of the Klebsiella isolates, 91% were K. pneumoniae, and the remaining 9% comprised K. ornithinolytica, K. oxytoca and K. terrigena. Ongoing molecular studies show that several of these also have ESBLs, though this was not evident in synergy studies, even using cefepime and cefpirome with clavulanate (M. J. Ellington, Antibiotic Resistance Monitoring and Reference Laboratory, Centre for Infections, personal communication). MIC distributions of temocillin for β-lactamase-producing Enterobacteriaceaea Modal values are indicated in boldface. In addition to the species groups detailed, this total also includes two Hafnia spp. and one Pantoea sp. with hyperproduced AmpC enzymes and one Proteus mirabilis with an ESBL. Includes 13 isolates with acquired CIT types. Includes 79 isolates representative of strain A, also detailed separately below. Of the Klebsiella isolates, 91% were K. pneumoniae, and the remaining 9% comprised K. ornithinolytica, K. oxytoca and K. terrigena. Ongoing molecular studies show that several of these also have ESBLs, though this was not evident in synergy studies, even using cefepime and cefpirome with clavulanate (M. J. Ellington, Antibiotic Resistance Monitoring and Reference Laboratory, Centre for Infections, personal communication). Temocillin MICs were determined using the BSAC agar dilution method, except that those for the carbapenem-resistant K. pneumoniae were determined using Etest (AB Biodisk, Solna, Sweden) on Iso-Sensitest agar (Oxoid, Basingstoke, Hants, UK). MIC distributions for groups of survey isolates are shown in Table 1. Modal values for AmpC and ESBL producers for all major species groups were 8 mg/L, identical to the values for the E. coli ATCC 25922 control, which is ampicillin-susceptible, and to E. coli ATCC 35218, which has only a classical TEM-1 β-lactamase. More than 88% of the AmpC- and ESBL-producing survey isolates were susceptible to temocillin at ≤16 mg/L and 99% at ≤32 mg/L. The susceptibility distribution for E. coli strain A, a CTX-M-15 β-lactamase-producing clone that is nationally widespread and locally dominant,7 mirrored this pattern: 79 strain A isolates were included, and, again, the modal MIC was 8 mg/L, with just three requiring MICs of 32 mg/L. Among the 13 ertapenem-resistant K. pneumoniae, 11 required temocillin MICs of 32 or 64 mg/L; the remaining two were more susceptible, suggesting that, as with the carbapenems, temocillin is vulnerable to the combination of ESBL and impermeability present in these organisms. It remains to relate the temocillin MICs to attainable levels and outcome data in the absence of a BSAC or EUCAST breakpoint. One recent open-label Belgian study describes clinically successful use of temocillin in 25 of 26 patients with urinary infections caused by multiresistant Enterobacter aerogenes, of which most were ESBL producers.8 Temocillin achieves urinary levels of 400–600 mg/L, suggesting a considerable margin over the MICs.9 More generally, a susceptibility breakpoint of ≤16 mg/L is used for systemic isolates in Belgium, based on a Cmax of 160 mg/L and t1/2 of 5 h from a 1 g intravenous dose, along with 85% serum binding.9 This indicates quite a narrow therapeutic margin for Enterobacteriaceae—whether ESBL-producing or not—outside the urinary tract. Nevertheless, six of the seven patients with sepsis caused by multiresistant E. aerogenes were cured in the Belgian study,8 although its authors note that these were cases with ‘less severe’ sepsis, and that carbapenems were preferred for more severe cases. We conclude that temocillin retains good in vitro activity against most current isolates of Enterobacteriaceae with ESBLs or hyperproduced or acquired AmpC, and also against K. oxytoca with hyperproduced K1 enzyme. It is a potential alternative to carbapenems against the growing number of infections, particularly those of the urinary tract, that are caused by ESBL producers and other cephalosporin-resistant strains. None to declare. We are grateful to Belpharma for financial support. 1Antibiotic Resistance Monitoring and Reference Laboratory, Centre for Infections, Health Protection Agency, 61 Colindale Avenue, London NW9 5EQ, UK; 2Healthcare-Associated Infections and Antimicrobial Resistance Department, Centre for Infections, Health Protection Agency, 61 Colindale Avenue, London NW9 5EQ, UK