| Abstract |
Background: Antimicrobial resistance has been increasingly reported worldwide and is associated with significant burden, including higher mortality, prolonged hospitalizations, and increased costs. Furthermore, increasing antimicrobial consumption of broader and broader-spectrum antimicrobials, combined with a slow antimicrobial pipeline, has resulted in a spiral of increasing resistance and emergence of extensively drug-resistant (XDR) and pandrug-resistant (PDR) Gram-negative bacteria (GNB), for which treatment options are extremely limited and the relevant literature scarce.
Aims of the thesis: Considering the emerging problem of XDR/PDR GNB, especially in Greek hospitals, works conducted as part of this thesis aim to review the epidemiology of XDR/PDR GNB, prognosis of infections caused by these bacteria and treatment options. During the course of this thesis, most works were eventually focused on PDR A. baumannii (PDRAB), which has emerged as a major pathogen in hospital-acquired infections and is difficult to treat.
Outline and findings of the thesis:
Systematic literature review regarding PDR GNB: At the beginning of this work, literature on PDR GNB was scarce and scattered. Therefore, a systematic review was first conducted to consolidate the relevant literature and to identify research gaps and guide further work on the topic. This review revealed a worldwide (25 countries, 5 continents) spread of PDR GNB, predominantly A. baumannii, K. pneumoniae and P. aeruginosa, with many reports coming from Greece. The potential for intra- and inter-institutional spread, as well as international dissemination was highlighted. Most affected patients were critically-ill and all-cause mortality was high, but the spectrum of excess mortality attributable to infection was unclear. Use of in vitro active antimicrobial regimens (based on newer antimicrobials, increased exposure treatment regimens and antimicrobial combinations) was associated with better outcomes, but available evidence was at high risk of bias as it was based exclusively on case reports and small case-series.
Review of treatment options for PDR GNB: Based on additional literature review treatment algorithms were proposed for the treatment of carbapenem-resistant K. pneumoniae, P. aeruginosa and A. baumannii that are co-resistant to other last resort options (aminoglycosides, polymyxins, tigecycline), by considering the different mechanisms of resistance to carbapenems of these pathogens. Antimicrobial combinations were identified as the only currently available option against PDRAB, but available evidence is limited to in vitro studies, animal models or case reports and small series. Although a vast number of potential combinations have been proposed in the literature, reported synergy was found frequently to be infeasible in vivo, highlighting the need to reconsider the methodology for identifying clinically relevant antimicrobial combinations against PDR GNB.
Estimation of excess mortality attributable to PDRAB: Considering uncertainty in the literature regarding the magnitude of mortality attributable to the infection itself rather than underlying comorbidities, a 4-year cohort study was conducted comparing patients infected by PDRAB to patients colonized by PDRAB (no-infection counterfactual model). The PDRAB colonization group was used to approximate non-infection-related mortality. The two groups had similar baseline characteristics, but the absolute excess risk of 30-day mortality in infected patients was 34%. Furthermore, multivariable competing risks regression showed that PDRAB infection significantly increased the daily hazard of 30-day in-hospital death while simultaneously leading to longer hospitalization.
Epidemiological clues for a transient PDRAB phenotype: A systematic review of the literature identified several reports of in-vivo-emergent (during treatment) resistance of A. baumannii to last resort antimicrobials (colistin and tigecycline), often associated with treatment failure. Furthermore, a high prevalence of colistin heteroresistance was estimated based on a systematic review and meta-analysis. However, emergent resistance was often associated with fitness cost, potentially resulting in a transient PDRAB phenotype emerging during antimicrobial selection pressure but being outcompeted by fitter XDRAB after withdrawal of antibiotics. To assess this, the antimicrobial resistance phenotype of sequential A. baumannii isolates from single patients from the above cohort were analyzed. Isolation of PDRAB was frequently preceded by isolation of XDRAB (suggesting emergence within the patient of the PDRAB, presumably due to antimicrobial selection pressure), and/or followed by isolation of XDRAB (an indication that the pandrug-resistance phenotype may be associated with fitness cost compared to the XDR phenotype). However, persistence of the PDR phenotype was demonstrated in 2 cases in the absence of continuous antimicrobial selection pressure suggesting the potential for emergence of fitter and more sustainable PDRAB strains.
Frequency and impact of polymicrobial A. baumannii infection: In the above-mentioned cohort, PDRAB was co-isolated with other bacteria in about a third of the cases. Assessing the impact of polymicrobial (vs monomicrobial) A. baumannii infection on outcomes has important implications for the design and analysis of clinical studies of A. baumannii infections. A systematic review and meta-analysis was conducted to estimate the proportion of infections that were polymicrobial and the magnitude of the association between polymicrobial (vs monomicrobial) infection and mortality. About 1 of every 3 A. baumannii bloodstream and respiratory tract infections were found to be polymicrobial, although considerable heterogeneity was observed. Polymicrobial A. baumannii infections was associated with lower 28-day mortality, suggesting that co-pathogens, which are usually more susceptible and more likely to be covered with effective antimicrobials, may represent the true pathogen in some polymicrobial ABC infections. However, the level of evidence was GRADEd as very low.
Conclusion: PDR GNB are increasingly being reported worldwide and are associated with significant mortality and limited treatment options. Especially problematic in Greek hospitals is A. baumannii, which has accumulated several mechanisms of resistance and can rapidly develop resistance in vivo under antimicrobial pressure. Although pandrug resistance may be associated with fitness cost, stable PDRAB strains can eventually emerge. More effective antimicrobial regimens (new antimicrobials, optimized increased exposure treatment regimens and antimicrobial combinations) could improve outcomes in infections by PDR GNB. Concerning infections caused by PDRAB, antimicrobial combinations are the only currently available option, but evidence is lacking and in vitro synergy (as currently defined) is often clinically irrelevant. Therefore, research to identify optimal treatment regimens is necessary, but emphasis on infection control and prevention and antimicrobial stewardship is ever more important.
|
Search
