Guest post: Patrick Harris, Staff Specialist in Microbiology, Central Laboratory, Pathology Queensland, Brisbane
Infectious disease and microbiology trainees diligently learn about organisms that have a type of broad-spectrum beta-lactamase (i.e. an enzyme that can inactivate betalactam antibiotics) called ‘AmpC’. These enzymes are encoded by a gene that is usually located on the chromosome of many Gram-negative enteric organisms. In most species (like E. coli) the ampC is not expressed to any significant level and does not cause much of a problem. However, some species – especially Enterobacter spp., Serratia marcescens, Citrobacter freundii and Morganella morganii (the ‘ESCPM’ group) – have an additional genetic apparatus that allows ‘inducible’ activity of the ampC gene when the bacteria are exposed to betalactam antibiotics. In such a state, the expression of AmpC increases to the extent that it inactivates antibiotics like amoxicillin or cephazolin and – because clavulanate does not inhibit AmpC – also amoxicillin+clavulanate. This explains why the ESCPM species are intrinsically resistant to these antibiotics. However, if you remove the drug exposure, the expression of AmpC returns to a low level.
(Semin Respir Crit Care Med 2015; 36(01): 56-73)
When ‘third generation’ cephalosporins (like ceftriaxone) became widely used in the late 1980s, a further problem emerged. Species like Enterobacter cloacae will often test susceptible in vitro to such antibiotics. But, patients with serious infections (like bacteraemia) would fail therapy if treated with these drugs, and the organism would appear to develop resistance while on treatment. In a classic study from 1991, this was observed in 6 of 31 (19%) patients with E. cloacae infections treated with a third generation cephalosporin (3GC). The problem was found to lie in mutations occurring in the genetic mechanisms that usually control the expression of ampC – when exposed to broad spectrum antibiotics, sub-populations of these more resistant ‘mutants’ expressing very high levels of the enzyme were selected and were able to resist therapy. Hence there has always been an understandable reluctance to trust beta-lactams to treat these types of bacteria. This has often forced clinicians to treat with carbapenems or quinolones – with downstream problems for antimicrobial stewardship.
In Australian laboratories it has been common practice to extend this natural caution about using 3GCs against Enterobacter to other beta-lactam agents, such as piperacillin+tazobactam or even cefepime. Furthermore, these ‘rules’ are often extrapolated to other AmpC-producers . However, the clinical evidence for this is shaky. Firstly, not all AmpC-producers are the same – Enterobacter cloacae produce more than 100 times the level of AmpC than, say, a Serratia species. The risk of clinical failure with 3GCs has only ever been replicated for Enterobacter infections – it is seemingly rare in other AmpC producers. Although piperacillin+tazobactam may become ineffective in the face of high-level AmpC production, it is probably less likely to select for ‘hyper-producing’ variants. Furthermore, there is almost no robust clinical data to suggest that the drug is clearly inferior in this context.
We undertook a systematic review and meta-analysis of any studies we could find that compared carbapenems with alternative antibiotics for the treatment of bloodstream infections caused by AmpC producers, including some local Hunter New England data. Firstly, it is worth noting that the overall quality of studies was relatively poor – with no RCTs, for instance, and most being small single-centre retrospective cohort studies. Nevertheless, we could find no clear evidence that carbapenems were a superior choice when compared to alternatives, such as piperacillin+tazobactam or cefepime in terms of 30-day mortality. Patients given quinolones seemed to have a better mortality outcome, but when adjusted for co-morbidity, this effect was eliminated suggesting that quinolones (being orally available) may be given in less sick patients rather than being a more effective option. (J Antimicrob Chemother (2016) 71 (2): 296-306)
We need better evidence to make informed decisions in response to these difficult infections – ideally in the form of RCTs. However, when you are tempted to automatically reach for meropenem to treat an ESCPM organism, ask yourself: is it really necessary? For many circumstances (e.g. urinary tract infections, abdominal infection with adequate source control) an alternative agent that tests susceptible in vitro may be just as adequate (e.g. co-trimoxazole, piperacillin+tazobactam, cefepime).