Fosfomycin is an old antibiotic that has reemerged as a new strategy to overcome antibiotic resistance. Fosfomycin spectrum of action includes grampositive (Staphylo-coccus aureus, Enterococcus spp.) and gram-negative bacteria (Enterobacterales, Pseudomonas spp.), including multiresistant microorganisms such as extended-spectrum beta-lactamase (ESBL) and carbapenemase producers⁶.

Its mechanism of action is cytosolic inactivation of N-acetylglucosamine enolpyruvyl transferase (MurA), which prevents the formation of N-acetylmuramic acid from N-acetylglucosamine and phosphoenolpyruvate, being the initial step of peptidoglycan synthesis. Fosfomycin crosses the bacterial wall using the transport system glycerol-3-phosphate (GlpT) and the transport system for hexose phosphates (UhpT)6.

There are two presentations of this antibiotic, fosfomycin disodium (for parenteral use) and fosfomycin trometamol (FT) with good oral bioavailability and only indicated for lower UTI8.

The increase in the clinical use of fosfomycin has led to the development of different resistance mechanisms, par-ticularly, when used as monotherapy8.

Fosfomycin resistance can result from reduced perme-ability (mutations in the chromosomal glpTand uhpTgenes), amino acid mutations in the active site of the MurA target, and the production of fosfomycin-inactivating enzymes (fos genes). Different fosfomycin-modifying enzymes have been described. FosA (glutathione S-transferase), the first to be described, is a metalloenzyme transferred through plasmids in Enterobacterales. The FosA3 enzyme is the most commonly identified FosA-like determinant found as an acquired mechanism in Escherichia coli6. The fosA3 gene is usually located in conjugative plasmids also carrying CTX-M-type ESBL-encoding genes8, and related to different plasmid incompatibility groups (IncN, IncF, IncFII, IncI1, among others)2. Other fos genes, such as fosA4, fosA5, and fosA6 have also been identified in E. coli, although less frequently8. In some Enterobacterales (i.e, Serratia marcescens, Klebsiella spp., Enterobacter spp., Kluyvera georgiana) there are homologous chromoso-mal fosA genes and these isolates can present reduced fosfomycin susceptibility6. Other fosfomycin-modifying enzymes are FosB (in gram-positive bacteria), FosX (chromosomal enzyme in Listeria monocytogenes) and FosC (in Pseudomonas syringae). Kinases that cause fosfomycin degradation through phosphorylation (FomA and FomB) have been identified in Streptomyces spp8.

In relation to the clinical use of fosfomycin, although it is an old antibiotic, the experience in children is scarce. Recently, a review of fosfomycin use in pedi-atrics recommends it for UTI treatment, osteomyelitis (such as in cases caused by methicillin-resistant S. aureus), and in combination for multidrug-resistant gram-negative bacilli in bacteremias (especially enterobacteria harboring carbapenemases)1. In the last years, FT was introduced as an option for the treatment of lower UTI in children older than 6 years of age and adults12.

Despite the fact that fosfomycin use is increasing, suscep-tibility to FT is not routinely tested in all laboratories and there is sparse data available about susceptibility profiles in E. coli isolates from UTI in children. Hereby we propose to describe the antibiotic susceptibility of isolates obtained from urinary tract infections in children older than 6 years old, and to characterize FT resistance mechanisms involved in resistant strains.

Children aged 6-14 years old with UTI, assisted at a hospital in Uruguay during 5/15/2018-8/15/2019, were included. UTI was defined as the presence of suggestive UTI clinical symptoms (dysuria, increased frequency of uri-nation, fever) and confirmed by a significant bacterial count (1 x 105CFU/ml) and monomicrobial growth in urine culture. Urine culture, bacterial identification and antibiotic susceptibility were assessed by automated systems either Vitek®2 (bioMerieux, Marcyl’Etoile, Francia, ASTN250) or Phoenix® (Becton Dickinson Diagnostics, Sparks, MD, EEUU, panel BGN 448879) and results were interpreted according to CLSI.4

Susceptibility to fosfomycin was tested by disk diffusion (200ug disc OXOID®, Hampshire United Kingdom), in accor-dance with CLSI guidelines4. Unfortunately, CLSI breakpoints to fosfomycin only apply to E. coli urinary tract isolates, which makes the study difficult for other Enterobacterales. In this work, we employed the susceptible E. coli breakpoint >16 mm for fosfomycin to all Enterobacterales. For resistant isolates, minimum inhibitory concentration was determined (MIC) by agar dilution, using Muller Hinton agar (OXOID®, Hampshire United Kingdom) supplemented with glucose-6-phosphate (25mg/l)4.

Polymerase chain reaction (PCR) for fosA and fosA3 was performed in Fosfomycin-resistant isolates10. Isolates harboring a fosfomycin resistant gene were further characterized by PCR for extended spectrum beta-lactamases.

Table 1: Antibiotic susceptibility results for Enterobacterales (data exposed as susceptible to each antibiotic). Percentages are expressed for E. coli isolates. 

Positive results were confirmed by Sanger sequencing, and sequences were analyzed using the BLASTn database (https://blast.ncbi.nlm.nih.gov).

Their phylogenetic group was characterized according to Clermont et al.3, and sequence type (ST) classification was done following the scheme to identify major E. coli ST caus-ing urinary tract and bloodstream infections (ST69, 73, 95, and 131) proposed by Doumith et al7.

Mobility of resistance genes was analyzed by conju-gation assays using rifampicin-resistant E. coli J53-2 as recipient9. Transconjugants were selected on Luria Bertani agar plates supplemented with fosfomycin (1mg/l) and glucose-6-phosphate (25mg/l) plus rifampicin (150mg/l).

Incompatibility groups and toxin-antitoxin mechanisms (i.e.: addiction systems) were detected by PCR using donor and transconjugant genomic DNA as template as reported previously9.

Plasmid size was estimated in donor strain and transcon-jugant by treatment with S1 nuclease followed by pulsed-field gel electrophoresis (PFGE)9.

Clinical data of patients with UTI caused by fosfomycin-resistant bacteria were collected by revision of medical records. Variables collected were age, sex, symptoms and signs, urine exam results, whether it was the first episode of UTI or recurrence, presence of morphological or functional abnormalities of the urinary tract and antibiotic consump-tion in the last 6 months. The study was approved by the ethics committee and board of directors.

From May 2018 to August 2019, 286 urine cultures with significant growth (1 x 105CFU/ml) were obtained. E. coli was the most frequent etiological agent in 255 urine cultures (89.2%), followed by P mirabilis in 14 (4.9%), K. pneumoniae in 7 (2.5%), Staphylococcus saprophyticus in 5 (1.7%), and others in 5 (1.7%).

Antibiotic susceptibility for the total of isolates was (susceptible results): 280 to fosfomycin (97.9%), 158 to ampicillin (55.2%), 265 to amoxicillin-clavulanate (92.7%), 279 to cefazolin (97.6%), 283 to cefuroxime and ceftriax-one (99%), 266 to ciprofloxacin (93%), 270 to nitrofurantoin (94.4%), 215 to trimethoprim-sulfamethoxazole (75.2%) and 279 to gentamicin (97.6%). Six isolates were resistant to fosfomycin (5 S. saprophyticus with fosfomycin natural resistance and 1 E. coli). Antibiotic susceptibility results for main enterobacteria species are shown in Table 1. E. coli showed high susceptibility levels (above 90%) to amoxicillin-clavulanate, cephalosporins, fluoroquinolones, aminoglycosides, nitrofurantoin and fosfomycin.

ESBLs were phenotypically detected in 2 E. coli isolates, but only one isolate (Ec870) was resistant to (Table 1 and 2). Ec870 was obtained from the urine of a 6-year old girl with dysuria and abdominal pain. This was her fifth UTI episode in the last 2 years, all caused by E. coli isolates with the same resistance profile (only resistant to ampi-cillin and trimethoprim-sulfamethoxazole). In the last 6 months she had received multiple antibiotic treatments with amoxicillin-clavulanate and cefuroxime. No morphological or functional urinary tract abnormalities were detected.

Ec870 harbored the ESBL gene blaCTX-M-14, which explains the resistant phenotype to ampicillin, cefazolin, cefuroxime and ceftriaxone, and the fosfomycin resistance gene fosA3 responsible for fosfomycin resistance (Table 2). The isolate belonged to phylogenetic group D and ST69. Two plasmids of approximately 45 kb and 150 kb in size, incompatibility groups F, FII, FIA, FIB and N, and addiction systems vagCD, hok/sok, ccdAB, were detected in Ec870. Conjugation assays were positive, and blaCTX-M-i4/fosA3 were co-transferred and detected in transconjugants (TcEc870). TcE870 presented only a 45 kb plasmid, IncN type. None of the toxin-antitoxin genes studied were detected in transconjugants. MIC to fosfomycin was 256mg/l for donor and transconjugant strains.

Community-acquired UTI caused by antibiotic-resistant bacteria is increasing worldwide, leading to the need for alternative antibiotic plans. Even though fosfomycin is an old antibiotic, FT was introduced in the last years for the treatment of lower UTI in children older than 6 years old. Fosfomycin susceptibility is not always tested in clinical microbiology laboratories, and scarce data are available in children.

In this study, E. coli was the predominant etiological agent (89.2%) of UTI in children aged 6-14 years, followed by P mirabilis (4.9%) and K. pneumoniae (2.5%). The dis-tribution of the etiological agents in the same age group in Spain is similar to that reported here14.

Table 2: Phenotypic and genotypic characteristics of fosfomycin-resistant Ec870. 

With regard to antibiotic susceptibility, 97.9% were susceptible to fosfomycin (resistance was detected in five S. saprophyticus with natural resistance and one E. coli with acquired resistance to fosfomycin). In Spain, susceptibility to fosfomycin from UTI isolates in children aged 5 -15 years was 88%, most of the resistant isolates corresponded to K. pneumoniae and only 2% of E. coli isolates were resistant to fosfomycin14.

In this work, we identified one conjugative plasmid from the IncN family harboring fosA3 and blaCTX-M-i4 genes. IncN plasmids harboring blaCTX-M-14 were reported in Salmonella spp. strains in our country5. However, this is not only the first report of fosA3 but also the association of blaCTX-M-14/fosA3 in an E. coli isolate in Uruguay.

Although E. coli ST69, harboring blaCTX-M-9 (another mem-ber of the blaCTX-M-9-group along with blaCTX-M-14) has been previously detected in adults in Uruguay15, its isolation in children is novel. Contrary to these results, commonly, ST69 isolates are considered susceptible to almost all the antibi-otic families7. As far as we know, this is the first detection of fosA3/blaCTX-M-14 in a ST69 E. coli isolate of human origin worldwide.

In this sense, the occurrence of these resistance deter-minants in a conjugative plasmid brings the possibility to transfer to other sequence types with resistance to more antibiotic families such as ST1317.

On the other hand, this finding is worrisome due to the possible dissemination of these resistance mechanisms to susceptible strains. Considering this, the UTI treatment with second or third-generation cephalosporins or fosfomycin tromethamine could exert selective pressure for the occur-rence of isolates resistant to both antibiotic families, as such was the case with this patient who had received cefuroxime for the treatment of previous UTI episodes.

Most of the strains studied in this work showed high susceptibility to different antibiotics, except ampicillin and trimethoprim-sulfamethoxazole in which resistance exceeded 20%. In a cohort of children in Spain, resis-tance rates to ampicillin, amoxicillin-clavulanate and trimethoprim-sulfamethoxazole were above 20%14. Our local results showed high susceptibility rates to amoxicillin-clavulanate, endorsing its use for oral UTI empirical treatment. In the cohort of children included in our study, ESBL enterobacteria were detected in only 2 urine cultures (0.7%), a frequency that was lower than that reported in France (0.8-10%)11 and Spain (1.8%)14.

The high susceptibility to many oral antimicrobial agents in our setting broadens empirical treatment options. For lower UTI, amoxicillin-clavulanate and nitrofurantoin could be suitable options. Fosfomycin shows several advan-tages for its use, such as once-daily dosing, low side effects, suitable clinical and microbiology results, and lit-tle effect on intestinal microbiota13. However, considering that fosfomycin is active against ESBL and carbapenemase enterobacteria producers, its use is not recommended when there are other treatment options of a narrower spectrum. As its use increases, active monitoring of resistance levels will be necessary.

Conflict of interest

The authors declare that they have no conflicts of interest.