Randomized-controlled Trial (RCT) on Combination Antibiotic for Infections Caused by Gram-negative Bacteria

Antimicrobial resistance is a global public health threat and the theme of the World Health Day 2011. While the issue in most cases (such as extensively-drug-resistant tuberculosis, antiviral-resistant human immunodeficiency virus and drug-resistant malaria) is the access to effective antimicrobial agents and/or the high cost of these drugs, for a small but increasing number of nosocomial drug-resistant Gram-negative bacteria, there is no safe and effective antibiotic available - not now nor in the next 10-year horizon, given the timeline of drug development. Extensively-drug-resistant Gram-negative bacilli (XDR-GNB) include the majority of the six organisms on the Infectious Disease Society of America's (IDSA's) watch list of global "bad bugs" for which the development of new drugs was urgently required. These Gram-negative bacilli are Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Acinetobacter baumannii. The most frequently isolated of these organisms in the Singapore hospital setting is XDR-A. baumannii at this point in time.

1. baumannii has swiftly emerged over the past three decades as a major nosocomial opportunistic pathogen, causing infections in debilitated patients especially in the intensive care unit (ICU) setting. It is the 10th most frequently isolated pathogen in US hospitals but ranks among the top 5 pathogens in tropical hospitals including Singapore, where the number of antibiotic-resistant A. baumannii infections exceeded 700 cases in 2006. The major issue with A. baumannii is that the organism has rapidly developed resistance to many antibiotics - as an example, within a single decade (1995 to 2004) in US hospitals, carbapenem resistance in A. baumannii increased from 9% to 40%. XDR-A. baumannii (defined as A. baumannii resistant to all antibiotics - including carbapenems, beta-lactam/beta-lactamase-inhibitors, cephalosporins, aminoglycosides, fluoroquinolones, tetracyclines and sulphonamides - with the exception of the polymyxins and tigecycline) has now been described causing infections in hospitals worldwide. Mortality associated with severe infections caused by multidrug-resistant A. baumannii and other XDR-GNB has ranged between 30% and 70% depending on the clinical setting and condition of the patients. Locally, the mortality from severe XDR-GNB infections is approximately 40%. From 2006 to 2010, there was an average of 140 cases of severe XDR-GNB infections in local hospitals each year, translating to approximately 56 deaths from infections attributable to this organism alone in Singapore every year. Other local XDR-GNB includes XDR-P. aeruginosa, and carbapenemase-producing E. coli and K. pneumoniae (carriage of New Delhi metallo-beta-lactmase-1 (NDM-1), OXA-48 and Klebsiella pneumoniae carbapnemase (KPC) genes). These remain relatively rare in Singapore, with fewer than 30 severe infections in local hospitals each year.

Treatment of infections caused by XDR-GNB presents a considerable challenge for clinicians. Monotherapy polymyxins - commercially available as polymyxin B or polymyxin E (colistin) - are currently the gold standard of treatment for severe XDR-GNB infections. However, they are associated with significantly more adverse effects and may be less effective clinically compared to other antibiotics, such as the beta-lactams. In one tertiary center in Korea, retrospective analysis suggested that mortality of XDR-A. baumannii bacteremia was not reduced when colistin was used for treatment compared to other antibiotics (to which the organisms were resistant). Individual treatment failures with polymyxins have been reported, either due to the development of resistance in vivo or inherent heterogeneous polymyxin resistance - a phenomenon where many isolates that appear susceptible to the drug may actually harbor polymyxin-resistant subpopulations. Tigecycline susceptibility in XDR-GNB is variable and clinical failures have also been reported, particularly for bloodstream infections due to the low achievable concentrations of the drug in serum as well as the potential for development of resistance during treatment. This has led to some experts advocating combination antibiotic therapy as an alternative.

In general, combination antibiotics have performed better than single agent polymyxin B in in vitro time-kill studies and animal models of infection. In accordance with other published reports, we have also shown that various antibiotic combinations demonstrated synergistic activity against XDR-GNB. The most effective in vitro combinations for local XDR-GNB isolates have been polymyxin B + rifampicin and polymyxin B + doripenem, with additive/synergistic effect in up to 50% of isolates without antagonism seen in the other isolates. However, it is uncertain if in vitro results in this particular instance directly predict clinical outcomes. No rigorous clinical trials have been completed to date and existing results based on case series and retrospective reviews are conflicting. Against local XDR-P. aeruginosa, however, dual antibiotic therapy appears to be less promising, with synergism achieved only when triple antibiotic combinations were tested.

Other in vitro studies have suggested that triple antibiotic combinations may be more effective than dual antibiotic combinations. Nonetheless, this is difficult to recommend in clinical practice at the current time because of the very probable rise in adverse effects versus uncertain benefits.

Because of the increasing number of XDR-GNB infections locally and worldwide, the questionable efficacy of the current gold standard monotherapy treatment, the paucity of novel and effective antibiotics against such infections for the foreseeable 10-year horizon, and the consistent reports of superiority in in vitro studies, it is critically important that combination therapy should be tested against polymyxin monotherapy in a rigorous clinical trial to ascertain if it represents a more effective treatment strategy. It is also important to determine if the results of multiple combinations bactericidal testing - like more standardized susceptibility testing for single antibiotics - will correlate well with clinical outcomes.