Activity of fosfomycin against nosocomial multiresistant bacterial pathogens from Croatia: a multicentric study

Aim To determine in vitro susceptibility of multiresistant bacterial isolates to fosfomycin. Methods In this prospective in vitro study (local non-random sample, level of evidence 3), 288 consecutively collected multiresistant bacterial isolates from seven medical centers in Croatia were tested from February 2014 until October 2016 for susceptibility to fosfomycin and other antibiotics according to Clinical and Laboratory Standards Institute methodology. Susceptibility to fosfomycin was determined by agar dilution method, while disc diffusion were performed for in vitro testing of other antibiotics. Polymerase chain reaction and sequencing was performed for the majority of extended spectrum β-lactamase (ESBL)-producing Klebsiella pneumoniae (K. pneumoniae) and carbapenem-resistant isolates. Results The majority of 288 multiresistant bacterial isolates (82.6%) were susceptible to fosfomycin. The 236 multiresistant Gram-negative isolates showed excellent susceptibility to fosfomycin. Susceptibility rates were as follows: Escherichia coli ESBL 97%, K. pneumoniae ESBL 80%, Enterobacter species 85.7%, Citrobacter freundii 100%, Proteus mirabilis 93%, and Pseudomonas aeruginosa 60%. Of the 52 multiresistant Gram-positive isolates, methicillin-resistant Staphylococcus aureus showed excellent susceptibility to fosfomycin (94.4%) and vancomycin-resistant enterococcus showed low susceptibility to fosfomycin (31%). Polymerase chain reaction analysis of 36/50 ESBL-producing K. pneumoniae isolates showed that majority of isolates had CTX-M-15 beta lactamase (27/36) preceded by ISEcp insertion sequence. All carbapenem-resistant Enterobacter and Citrobacter isolates had blaVIM-1 metallo-beta-lactamase gene. Conclusion With the best in vitro activity among the tested antibiotics, fosfomycin could be an effective treatment option for infections caused by multiresistant Gram-negative and Gram-positive bacterial strains in the hospital setting.

Fosfomycin is a phosphonic acid derivative with a broadspectrum antibacterial activity. It inhibits peptidoglycan assembly, thereby disrupting bacterial cell wall synthesis (1). It has been clinically available for decades and a single dose of fosfomycin has been widely accepted as the first line treatment for uncomplicated urinary tract infections (UTIs) (2). In the form of its trometamol salt, approximately 40% of the drug is absorbed following oral administration. After being released from trometamol by hydrolysis, fosfomycin is rapidly excreted unchanged by glomerular filtration and reaches high-peak urinary concentration of approximately 4000 mg/L. Fosfomycin treatment achieves comparable clinical and microbiological cure rates to longer courses of antibiotic alternatives (quinolones, beta lactams, aminoglycosides, nitrofurantoin, and sulfonamides) (3). Moreover, a number of in vitro studies have demonstrated that fosfomycin has an excellent activity against many multiresistant bacteria, including extended spectrum β-lactamase (ESBL) and plasmid-mediated ampicillin class C (AmpC)-producing Gram-negative bacilli and carbapenem-resistant Enterobacteriaceae (4,5). Based on the results of previous in vitro studies, we hypothesized that the majority of multiresistant Gram-negative and Grampositive bacteria, including carbapenem-resistant isolates, would exhibit susceptibility to fosfomycin. Compared to the previous studies, we included a larger proportion of carbapenem-resistant isolates, which present a considerable challenge for clinicians. Since the literature data about multiresistant Citrobacter freundii (C. freundii) susceptibility to fosfomycin is scarce, we also tested a relatively large number of C. freundii isolates. Therefore, we tested fosfomycin's in vitro activity against multidrug resistant pathogens for which very limited antibiotic options are available.

MATERIALS AND METHoDS
In this prospective in vitro study, with the level of evidence 3, 288 multiresistant bacterial clinical isolates were tested from February 2014 until October 2016 in the Clinical Hospital Center Zagreb for in vitro susceptibility to fosfomycin and other antibiotics. Multiresistance was defined as resistance to at least three members of different antibiotic groups (6). Isolates were collected from six medical centers in Zagreb and one in Split (Clinical Hospital Center Zagreb, Croatian National Institute of Public Health, Clinic for Infectious Diseases "Dr. Fran Mihaljević, " Clinical Hospital "Sveti Duh, " "Andrija Štampar" Teaching Institute of Public Health, Polyclinic Breyer for Medical Biochemistry and Microbiology, Clinical Hospital Center Split). The majority of isolates (75%) originated from the University Hospital Cen-ter Zagreb. Non-copy clinical isolates were consecutively collected during the period from February 2014 until October 2016. While the vast majority of Gram-negative isolates were from urine, Gram-positive isolates were also collected from other clinical specimens such as blood cultures, stool, ascites, central venous catheters, tracheostomy, throat, and wound swabs. The following number of isolates were tested, all having multiresistant phenotype: Escherichia coli (E. coli) (n = 72), Klebsiella pneumoniae (K. pneumoniae) (n = 62), methicillin-resistant Staphylococcus aureus (MRSA) (n = 36), Enterobacter species (spp) (n = 35), Proteus mirabilis (P. mirabilis) (n = 28), C. freundii (n = 23), Pseudomonas aeruginosa (P. aeruginosa) (n = 16), and vancomycin-resistant enterococcus (VRE) (n = 16). All E. coli and K. pneumoniae isolates were ESBL producers, and some of the K. pneumoniae isolates were also carbapenem-resistant. The isolates were identified to the species level by conventional biochemical testing and matrix-assisted laser desorption/ionization time-of-flight.
Extended spectrum beta-lactamase production was confirmed by a double disk synergy test according to Jarlier (7). Deformation of the inhibition zones around cephalosporin disks toward central disk containing clavulanic acid represented a positive result. In vitro sensitivity of multiresistant bacterial strains to fosfomycin was tested according to Clinical and Laboratory Standards Institute methodology (8). Thus, minimum inhibitory concentration (MIC) susceptibility breakpoint for fosfomycin was defined as 64 mg/L. Intermediate susceptibility was defined as MIC 128 mg/L and resistance as MIC≥256 mg/L. Fosfomycin MIC breakpoints were determined by agar dilution method, as this is the reference method for in vitro sensitivity testing to fosfomycin. Antimicrobial susceptibility testing to other routinely used antibiotics for UTI was performed by disk diffusion method using commercially-prepared, fixed concentration, paper antibiotic disks, in accordance with Clinical and Laboratory Standards Institute criteria (8).
We determined our sample size based on similar in vitro susceptibility studies (16)(17)(18). Statistical processing was carried out using the statistical package IBM SPSS Statistics, version 23 (IBM, Armonk, NY, USA, license owned by the Croatian Institute of Public Health). χ 2 test was used to determine whether there was a significant difference between the susceptibility rates of multiresistant bacterial isolates to fosfomycin and other antibiotics. The level of significance was set at P < 0.05.
All of the 72 E. coli isolates were ESBL producers. Of those, 70 (97.2%) were susceptible to fosfomycin, with the remaining two intermediately susceptible (Table 1). No resistance to carbapenems was observed. The difference in susceptibility rates between fosfomycin and all other antibiotics, except carbapenems, was significant (χ 2 test, P = 0.001).
Among the Enterobacter spp. isolates, the susceptibility rate to fosfomycin was very high (30/35, 85.7%). Since the majority of isolates were carbapenem-resistant, they were susceptible only to colistin and fosfomycin (Table 1). Fosfomycin showed significantly better rates of in vitro activity than amikacin and all other antibiotics except colistin (χ 2 test, P = 0.001). Resistance to amoxicillin/clavulanic acid, cefazolin, and cefuroxime observed in all isolates is due to  the production of intrinsic, chromosomal AmpC beta-lactamase of Enterobacter spp.

Methicillin-resistant
Eighty four percent of multiresistant Gram-negative isolates were in vitro susceptible to fosfomycin and 69% were susceptible to carbapenems. Susceptibility rates to other antibiotic groups were much lower ( Table 6). The differ-ence between susceptibility rates to fosfomycin and carbapenems was significant (χ 2 test, P = 0.001).
We performed a polymerase chain reaction analysis of 36/50 ESBL producing K. pneumoniae isolates. The majority of isolates had CTX-M-15 beta lactamases and ISEcp insertion (27/36) sequence, which enhances gene expression and level of resistance, and is important for the gene mobilization (20). All of the carbapenem-resistant Enterobacter and Citrobacter isolates were carrying bla VIM-1 metallo-betalactamase gene, while Citrobacter isolates also had chromosomaly-encoded CMY AmpC-type beta-lactamase. Of the 8 carbapenem-resistant K. pneumoniae isolates, one was oxacillinase-48 positive and the other had Verona integron-encoded metallo-β-lactamase-1 (Table 7).

DISCuSSIoN
The main finding of our study is high susceptibility rate to fosfomycin (82.6%) among 288 multiresistant isolates, which affirmed our hypothesis. Both the Gram-positive and Gram-negative isolates showed high rates of sensitivity to fosfomycin. An important finding of our study is that fosfomycin had significantly better in vitro activity against multiresistant Gram-negative isolates than carbapenems. Since Gram-negative pathogens are the most common causative agents of nosocomial bacterial infections, the results of our study have implications for empirical treatment of suspected bacterial infections in the hospital setting. This is in accordance with recent trends of the revival of old antibiotics such as colistin, fosfomycin, and nitrofurantoin, as these drugs still exhibit high in vitro activity against evolving multiresistant Gram-negative pathogens (21,22). While nitrofurantoin is only used clinically for the treatment of UTIs, both fosfomycin and colistin are also used in the treatment of other infections, with the advantages of fosfomycin being the additional coverage of Gram-positive pathogens and lack of nephrotoxicity (4). The more common use of fosfomycin and colistin in appropriate clinical setting represents one of the carbapenem-sparing strategies and is promoted by the leading experts in the field (23).
Both ESBL-producing E. coli and K. pneumoniae exhibited high susceptibility rates to fosfomycin, with the main difference being higher MICs in K. pneumoniae isolates. Among ESBL-producing E. coli isolates, fosfomycin demonstrated similar in vitro activity as carbapenems, which represent the first-line group of antibiotics for the treatment of infections caused by these pathogens (23). An acceptable alternative was nitrofurantoin, which, however, has more limited antibacterial spectrum than fosfomycin and carbapenems. Based on our results, the use of other antibiotics could not be routinely recommended. It needs to be emphasized that the two isolates with intermediate susceptibility to fosfomycin were obtained from patients previously treated with prolonged courses of fosfomycin owing to complicated UTIs. The emergence of fosfomycin resistance during prolonged treatment courses has been reported in the literature and is quoted as one of the main concerns regarding the clinical utility of fosfomycin (24). The susceptibility rate of ESBL-producing isolates to quinolones and trimethoprim-sulfamethoxazole was especially low, as expected since plasmids encoding ESBLs usually carry genes responsible for resistance to quinolones, sulphonamides, and aminoglycosides. The percentage of ESBL production is considerable, not only in hospitalized patients but also in outpatients. Based on the surveillance data from the Reference Centre for Antibiotic Resistance Surveillance of the Croatian Ministry of Health, 47% of K. pneumoniae and 13% of E. coli invasive isolates were ESBL producers in 2015, with a rise of ESBL production in E. coli compared to earlier years (25). Fosfomycin has  been widely accepted as the first-line antibiotic treatment of uncomplicated UTIs (1). The current Croatian national guidelines also recommend the use of fosfomycin as one of the treatment options for acute uncomplicated lower UTIs in women and for UTIs in pregnant women (26). Since the resistance to trimethoprim-sulfamethoxazole is high in E. coli and K. pneumoniae with and without ESBL production, fosfomycin and nitrofurantoin could be taken into the consideration as the first-line therapy for the UTIs caused by multiresistant pathogens.
Falagas et al (4) systematically reviewed 17 studies evaluating the antimicrobial activity of fosfomycin. Using a MIC susceptibility breakpoint of 64 mg/L, they found 1604 (96.8%) of 1657 ESBL-producing E. coli isolates and 608 (81.3%) of 748 ESBL-producing K. pneumoniae isolates to be susceptible to fosfomycin. We obtained similar results, with 97.2% and 80% of ESBL-producing E.coli and K. pneumoniae being susceptible to fosfomycin, respectively.
An unexpected finding is a high resistance to fosfomycin in carbapenem-resistant K. pneumoniae isolates (MIC 50 and MIC 90 128 and 512 mg/L, respectively). Although only 12 isolates were tested, MICs were higher than in VRE or P. aeruginosa isolates, which are expected to be more resistant. These isolates also had significantly lower susceptibility rate than ESBL-producing K. pneumoniae isolates. Since fosfomycin has a unique mechanism of action, it does not display cross-resistance with other antibiotics. Hence, this finding is not in concordance with the literature and should be confirmed on a larger number of isolates (27).
Our results show excellent fosfomycin activity against C. freundii isolates and are similar to those by Samonis et al (28), who found all 29 C. freundii isolates to be susceptible to fosfomycin. However, in the latter study, the majority of isolates were susceptible to cephalosporins and fluoroquinolones, and all but one isolate were susceptible to imipenem, while most of our isolates were resistant to carbapenems (28). Our results are similar to those by Hammerum et al (29), who tested 13 New Delhi metallo-beta-lactamase 1-producing C. freundii isolates, all of which were susceptible to fosfomycin.
Fifty-two of 73 isolates (71.2%) of carbapenem-resistant Enterobacteriaceae were susceptible to fosfomycin. The rate of susceptibility was higher than that reported by Livermore et al (30), who found 49/81 isolates (60.5%) to be in vitro sensitive to fosfomycin. The difference could be explained by the fact that their sample included more Klebsiella spp. isolates (64.2% of the whole sample), while ours included more Enterobacter spp. and C. freundii isolates, with lower MICs for fosfomycin (Enterobacter spp. and C. freundii accounted for 62% of all carbapenem-resistant isolates).
Fosfomycin activity against Enterobacter spp. was comparable to the results by Kaase et al (5). Pseudomonas isolates are known to have high MICs of fosfomycin. Fosfomycin had similar activity as amikacin and better activity than carbapenems, but the number of tested isolates was small.
Among Gram-positive isolates, we found excellent susceptibility rates for MRSA and low for VRE, which is in accordance with the data published earlier (3).
The main limitation of our study is the fact that, owing to financial restrictions, not all isolates were analyzed by pulsedfield gel electrophoresis. Hence, it cannot be excluded that some of the isolates were genetically identical and hence represent a clone. Moreover, the susceptibility of some isolates was not tested to all antibiotics. For example, susceptibility to amikacin was not tested in P. mirabilis, P. aeruginosa, and ESBL-producing E. coli and K. pneumoniae isolates.
According to our results, fosfomycin is a potentially effective treatment option for infections caused by multiresistant Gram-negative and Gram-positive bacterial isolates, with the exception of VRE. Fosfomycin demonstrated the best in vitro activity of all antibiotics tested against 288 multiresistant bacterial isolates. Hence, it is a valuable addition to antibiotic armamentarium in the hospital setting, especially for Gram-negative infections, in which antimicrobial resistance rates are rising and effective antibiotic options are scarce. In the context of the worrisome high rates of carbapenem-resistance among multiresistant Gram-negative isolates, fosfomycin represents a potentially valuable treatment option, both clinically and as a carbapenemsparing strategy. Declaration of authorship BL and BB designed the study, acquired, analyzed and interpreted the data, drafted and critically reviewed the manuscript. LR designed the study, analyzed and interpreted the data, drafted and critically reviewed the manuscript. EV, GMM, GA, and TI analyzed and