Gram-negative superbugs: inappropriate antimicrobial therapy and mortality

Keywords::

• antimicrobial resistance
• carbapenem resistance
• drug combinations
• inadequate therapy
• mortality

Antimicrobial resistance has and continues to complicate therapy for infectious diseases caused by drug resistance pathogens. Many have commented on and expressed concerns about the lack of new antimicrobials in the face of increasing antimicrobial resistance. Our sad reality is simply that resistance complicates therapy and may be responsible for clinical failure and mortality, resistance is increasing and newer therapies are lacking. This is not an encouraging picture.

‘Superbugs’ are those bacteria that are pathogenic and exhibit multidrug-resistant characteristics. Unfortunately, extended spectrum β-lactamase (ESBL)-producing Gram-negative bacilli join methicillin-resistant Staphylococcus aureus, vancomycin-resistant Enterococcus, vancomycin intermediate-and-resistant S. aureus and coagulase-negative Staphylococci with elevated vancomycin MIC as organisms of great concern to human health.

β-lactamase enzymes with Gram-negative bacilli are not new. Haemophilus influenzae and Moraxella catarrhalis β-lactamase-positive strains are prevalent, however while these organisms were resistant to ampicillin/amoxicillin, they were usually susceptible (99%) to second- and third-generation cephalosporins, fluoroquinolones and macrolides, as well as other drug classes. β-lactamase enzymes are also well recognized in Enterobacteriaceae and Pseudomonas aeruginosa. For many years, Gram-negative bacilli resistant to first generation cephalosporins remained susceptible to second- and third-generation cephalosporins and as such, alternatives for therapy within the β-lactam class were reasonable. The emergence of ESBL-producing Gram-negative bacilli rendering resistance against third-generation agents was most unwelcomed and compromised this important group of antibiotics. Extended spectrum β-lactam drugs factor prominently in published therapeutic guidelines for the treatment of moderate-to-severe infections including meningitis, intra-abdominal infection and nosocomial respiratory tract infection. Indeed, β-lactam agents may represent up to 60% of hospital formularies for antimicrobial use.

The world of β-lactamase enzymes is challenging and I suspect only a relative few in the world fully understand this most complicated area. To roughly differentiate between some of the extended spectrum β-lactamase enzymes, we can consider the following: ESBL enzymes hydrolyze all penicillins, cephalosporins and monobactams; ampC β-lactamase enzymes hydrolyze penicillins, cephalosporins (except cefepime – fourth generation) and monobactams (i.e., aztreonam); metallo-β-lactamase enzymes hydrolyze penicillins, cephalosporins, cephamycins and carbapenems; and Klebsiella pneumoniae carbapenemase enzymes hydrolyze penicillins, cephalosporins, cephamycins, carbapenems and monobactams. Quite simplistically, β-lactamase enzymes hydrolyze the β-lactam ring of the β-lactam molecule. Once hydrolyzed, the antibiotic is no longer biologically active and resistance results. Genes encoding for β-lactamase enzymes present on transmissible genetic elements (i.e., plasmid) facilities transmission between organisms making control difficult, if not impossible. Laboratories capable of distinguishing between these various enzymes add valuable information for clinicians as they help identify antimicrobials still useful for therapy as well as identifying compounds to which the pathogen is resistant. I would argue that the faster this information becomes available to clinicians, the greater the impact on patient care. A same-day result identifying drug resistance would be ideal – especially in those more critically ill.

ESBL organisms are showing up with increasing frequency in a number of clinical specimens. In our institution, ESBL-producing organisms have, to date, been seen most frequently in urinary tract infection. Unfortunately, isolates from respiratory tract infection and bacteremia are being seen with increasing frequency. Data from our institution are consistent with data summarized by Hyle et al. [1]. Treatment becomes more complicated as many ESBL strains are coresistant to other antimicrobials across other drug classes (i.e., aminoglycosides, trimetheprim/sulfamethoxazole and fluoroquinolones). Antimicrobials such as colistin and tigecycline may be useful in such cases. Macrolides play no role here as their activity is virtually nonexistent against Enterobacteriaceae.

Epidemiologically, ESBL-producing organisms have a worldwide distribution (with varying frequency) and considerable concern has been directed toward carbapenemase producing strains. Over the past 2–3 years, the strain New Delhi metallo (NDM-1) has attracted a tremendous amount of attention which, in my opinion, was justified owing to the initial concerns over pathogenicity and the pan-resistant profile of these strains. These strains have now been identified in many countries of the world and the initially isolated strains seem to have convincing epidemiological data tracing the origin back to India.

Of the β-lactam antibiotics, carbapenems (ertapenem, imipenem, meropenem) have remained important for therapy for carbapenem susceptible but β lactamase-producing organisms. Unfortunately, carbapenemase-producing strains have now compromized use of these agents. The carbapenems are not all the same. For Enterobacteriaceae, all compounds have good in vitro activity based on comparative MIC data against susceptible strains. Imipenem and meropenem have greater in vitro activity against P. aeruginosa than does ertapenem. Some suggest preferential use of ertapenem reduces selective pressures on the selection of carbapenem-resistant P. aeruginosa. Colistin has been used to treat multidrug-resistant strains but at the recent meeting of the European Congress for Clinical Microbiology and Infectious Diseases, held in Berlin, Germany on the 26–30 April 2013, colistin-resistant strains were a concerning observation of great discussion. Where do we go from here with multidrug-resistant Gram-negative bacilli and antimicrobial therapy? In the absence of new antimicrobials under development, combinations of drugs may require more extensive investigation. While drug combinations increase costs, the potential for drug interactions and also the potential for increased side effects, the options are limited.

There is no doubting the importance of clinical outcome and the impact of getting it right when choosing antimicrobial therapy. Rahal summarized data from a number of investigators comparing adequate and inadequate initial antibiotic therapy for intensive care unit (ICU)-acquired and ventilator-associated pneumonia [2]. Mortality was statistically significantly higher in patients receiving inadequate therapy. Similar observations were seen for inadequate antibiotic therapy in patients with bacteremia and severe sepsis/early septic shock – that is, mortality was higher than in patients receiving adequate therapy. Similarly, Hyle et al. commented on the impact of initial antimicrobial therapy on mortality in ESBL infections [1]. Inadequate initial antimicrobial therapy was an independent risk factor for mortality in nonurinary ESBL infection. Additional independent risk factors were multidrug-resistant ESBL and healthcare-acquired ESBL infections. The time to initiation of adequate antimicrobial therapy and mortality was significantly associated.

Kang et al. reported on bloodstream infections with ESBL-producing Escherichia coli and K. pneumoniae and investigated risk factors for mortality and also highlighted antimicrobial therapy [3]. The overall risk for 30-day mortality was approximately 26%; approximately 32% for those with K. pneumoniae and approximately 19% for those with E. coli. No significant difference was seen between a noncephalosporin and a broad-spectrum cephalosporin used empirically; however, a significant difference was seen when definitive antimicrobial therapy was compared between a noncephalosporin (~16% mortality) versus a broad-spectrum cephalosporin (~56% mortality; p < 0.001). Patients were also more likely (statistically significant) to die if they were neutropenic (~46% mortality), presented with septic shock (~80% mortality), received care in the ICU (~64% mortality), peritonitis (~68% mortality), immunosuppressive therapy (~54% mortality), corticosteroid therapy (~44% mortality) and an APACHE II score ≥16 (~88% mortality). Additionally of interest was the outcome of empiric treatment with a cephalosporin based on the MIC of the organism (i.e., E. coli or K. pneumoniae). Treatment failure (72 h after empiric therapy) was not seen in patients with strains having MIC ≤1 µg/ml, 25% (one out of four) for strains of 2 µg/ml and 67–11% for strains with MIC of 8–32 µg/ml. 30-day mortality paralleled the elevated MIC values with 27–50% 30-day mortality in patients with organism having MIC values of 8–32 µg/ml. For patients with strains having an MIC of 32 µg/ml, approximately 73% (eight out of 11) failed therapy and approximately 27% (three out of 11) died.

Carbapenem-resistant Enterobacteriaece are a major concern. Patel et al. reported on the outcomes of carbapenem-resistant K. pneumoniae and the impact of antimicrobial therapies [4]. The clinical risk factors associated (univariable analysis) with infection with a carbapenem-resistant organism (vs carbapenem-sensitive strain) included transplant recipient, use of central venous catheter, mechanical ventilation, ICU stay, prolonged length of stay before infection, prior Gram-negative antibiotic therapy, class of antibiotic used including cephalosporin, β-lactamase and/or β-lactamase inhibitor, carbapenem and aminoglycoside. Risk factors identified by multivariable analysis included transplant recipient, mechanical ventilation, prolonged length of stay before infection and use of a cephalosporin or carbapenem. Risk factors associated with mortality by univariable analysis included heart disease, renal insufficiency, ICU stay and adjunctive therapy (i.e., debridement, drainage, catheter removal). By multivariable analysis, risk factors for mortality included heart disease, ICU stay and adjunctive therapy as described.

To further complicate the ESBL story, Costa et al. reported on the detection of ESBL in E. coli strains from healthy pets in Portugal [5]. Similar observations were reported by O’Keefe et al. with E. coli urinary tract isolates from dogs and cats in the USA [6]. So, do we acquire ESBL-producing organisms from companion animals or do we give these organisms to our pets, and do we now serve as reservoirs to each other?

In a world of increasing prevalence of drug-resistant organisms, including ESBL-producing Gram-negative bacilli, what are our options? Clearly new drugs would be welcomed but not likely given the current drug development pipeline. Clinical microbiology diagnostic laboratories have an increasingly more important role to play. Previously, Dokouhaki and Blondeau commented on advances in technology in the microbiology laboratory and its potential impact on patient care [7]. So what is the role of the clinical microbiology laboratory? In today’s laboratories, technology exists to rapidly (in minutes) identify organisms once isolated, and susceptibility/resistance can be determined in ≤24 h. Specifically regarding ESBL-producing organisms, the lab has a responsibility to accurately and as quickly as possible confirm ESBL production and differentiate between carbapenem-susceptible and carbapenem-resistant strains. Such information has huge implications for patients with more severe infections, where multidrug-resistant phenotypes are seen and where a carbapenem would be the preferred agent alone or in combination. Optimization of therapy has been previously commented on as important for treatment [8]. To optimize therapy, consideration needs to be given to the clinical trial data available, the activity of the drug against the target pathogen(s) and the likelihood of the dose to satisfy the key pharmacological characteristics of the drug: in the case of β-lactam drugs, time above the MIC that is key for a time-dependent compound. Depending on the site of infection, serum versus tissue drug concentrations need consideration. Some data now suggest that continuous versus bolus infusion could extend the time above MIC [9]. Such information may have important consequences for optimizing therapy.

ESBL-producing stains are not going away. What might be a strategy going forward for therapy of Gram-negative infections? In addition to optimizing therapy when susceptible strains are encountered, there may be some value in re-exploring the value of combination therapy – not necessarily for synergistic effects but to reduce the likelihood for resistance. Minimal drug doses that exceed key pharmacokinetic/pharmacodynamic need to be explored. The value of this approach needs to be carefully re-examined and evaluated in a prospective approach. New treatment options (drugs or other treatment approaches) would be most welcomed, including novel approaches to drug delivery that would allow higher achievable and sustainable therapeutic drug levels without the systemic toxicities if possible. Some of the resistance data presented at the European Congress for Clinical Microbiology and Infectious Diseases this year clearly indicate we have an up hill battle.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.