Effectiveness of Fosfomycin for the Treatment of Multidrug-Resistant Escherichia coli Bacteremic Urinary Tract Infections

This randomized clinical trial investigates noninferiority of fosfomycin vs ceftriaxone or meropenem in achieving clinical and microbiological cure among patients with urinary tract infections due to multidrug-resistant Escherichia coli.


Introduction
Escherichia coli is one of the most frequently occurring human pathogens. After a massive use of cephalosporins and fluoroquinolones, multidrug-resistant (MDR) isolates have spread dramatically worldwide. 1,2 As a consequence, the consumption of last-resort drugs, such as carbapenems, increased over the last 20 years, 3 which in turn is facilitating the dramatic spread of carbapenem-resistance. 4 These outcomes suggest that finding alternatives for the treatment of MDR E coli infections is a medical need.
Some old drugs were inadequately developed according to present standards, which suggests that appropriate trials must be performed to evaluate the potential efficacy of these drugs. Targeted therapy is a potential indication for these drugs, allowing a decreased consumption of broadspectrum drugs. Fosfomycin, discovered more than 40 years ago, is active against a wide range of pathogens, including MDR Enterobacterales. 5,6 This drug is available for intravenous use as fosfomycin disodium in some countries (although not in the United States) and as an oral formulation (ie, fosfomycin trometamol). However, high-quality studies with fosfomycin are scarce. 5,7 Recently, it was shown to be noninferior to piperacillin-tazobactam for treatment of complicated urinary tract infections (cUTI). 8 Because cUTI includes highly heterogeneous infection types and considering that fosfomycin may be less efficacious against Enterobacterales than against other than E coli, 9,10 we conducted this study to test the hypothesis that fosfomycin is not inferior to ceftriaxone or meropenem for the targeted treatment of bacteremic UTI (bUTI) caused by MDR E coli.

The Fosfomycin vs Meropenem or Ceftriaxone in Bacteriemic Infections Caused by Multidrug
Resistance in E. Coli (FOREST) randomized clinical trial was conceived as a noninferiority trial intended to provide information on fosfomycin as an alternative drug to ceftriaxone and carbapenems, which are associated with increased risk of colonization and infection due to MDR bacteria; therefore, treatment with fosfomycin may have a protective effect for that risk. The Andalusian Ethics Committee approved this study, and written informed consent was obtained from all participants. The results are reported according to the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.  11 ; in January 2015, the protocol was modified owing to low recruitment to include any MDR E coli, and ceftriaxone was added as comparator for susceptible isolates. 12 The study protocol is available in Supplement 1.

Study Design and Patients
Hospitalized adult patients with monomicrobial bUTI due to E coli showing resistance to at least 1 drug from 3 different families to which wild-type E coli is susceptible 13 and susceptibility to fosfomycin and to ceftriaxone or meropenem were eligible if deemed to need at least 4 days of intravenous therapy. Exclusion criteria were septic shock, prostatitis, kidney transplantation, polycystic kidney disease, a more than 48-hour delay in abscess drainage or obstruction release, palliative care, heart failure New Yor Heart Association (NYHA) class III or IV, liver cirrhosis, hemodialysis, allergy to study drugs, and active empirical treatment for more than 72 hours at randomization.
Dose adjustments for patients with kidney dysfunction are specified in the study protocol (Supplement 1). After 4 days of intravenous treatment, a switch was allowed to an in vitro active oral drug. This was oral fosfomycin trometamol 3 g every 48 hours for patients assigned to fosfomycin and cefuroxime axetil, ciprofloxacin, amoxicillin-clavulanate, or trimethoprim-sulfamethoxazole at standard dosing for patients in the comparator group, according to the susceptibility profile of the isolate. To reflect real clinical practice, patients with ceftriaxone-resistant isolates in the comparator group could also be switched to parenteral ertapenem for ambulatory treatment. The recommended total duration of treatment was 10 to 14 days. The patients were followed up for 60 days.
Assignment to the treatment group was done centrally using a previously prepared list integrated in the electronic case report form. Randomization was stratified for empirical therapy (ie, active or not) and ceftriaxone susceptibility. No blocks were used. Investigators were not blinded for drug allocation, with the exception of 2 investigators (J.S.-D. and J.R.-B.) who were blinded for checking end points.

End Points, Study Populations, and Follow-up
The primary end point was clinical and microbiological cure (CMC) at 5 to 7 days after finalization of treatment (test of cure, TOC) in the modified intention-to-treat (MITT) population. 14 Clinical cure was defined as resolution of all new signs and symptoms of infection at TOC; microbiological cure was defined as no isolation of the causative E coli strain in blood cultures from day 5 or in urine culture at TOC. Clinical failure was defined as not reaching clinical cure at TOC, worsening signs or symptoms after 48 hours of treatment, or death. Microbiological failure was defined as isolation of E coli in blood culture at day 5 or in urine culture at TOC. Secondary end points included clinical and microbiological cure in the clinically evaluable population (CEP) and microbiologically evaluable population (MEP) at TOC, respectively; length of hospital stay; relapses (ie, reappearance of fever or UTI symptoms with isolation in blood or urine of E coli with Յ2 band differences in pulse-field gel electrophoresis [PFGE], or Յ2 drugs in susceptibility profile if not available for PFGE); reinfections (ie, the same categories as for relapses but with isolation of a different bacteria or E coli not fulfilling the previously mentioned criteria); 60-day mortality; and adverse events (AEs). Exploratory end points included blood levels of fosfomycin (already reported), 15

Microbiology and Rectal Carriage Substudy
Local microbiology laboratories used standard microbiological techniques for bacteria identification and susceptibility testing. Patients recruited at 3 hospitals were asked to participate in an exploratory substudy of rectal carriage by ceftriaxone-resistant or carbapenem-resistant Enterobacterales or Acinetobacter baumannii, using McConkey agar with cefotaxime (2 mg/L) or ChromID-ESBL (BioMérieux). Rectal swabs were taken at days 0, 3, or 4 and at end of treatment. All study isolates were sent to Hospital Universitario Virgen Macarena, where identification and antimicrobial susceptibility were confirmed using matrix-assisted laser desorption and ionization time of flight and microdilution, respectively. European Committee on Antimicrobial Susceptibility Testing recommendations 16 were used. ESBL and carbapenemase genes were characterized by polymerase chain reaction and sequencing, and clonality of isolates was studied by PFGE.

Data Monitoring
Collected data were verified with original data sources. Primary and secondary end points were

Statistical Analysis
To our knowledge, no previous trials on bUTI due to MDR E coli had been performed; we estimated a clinical cure rate of 85% with meropenem 17 or ceftriaxone 12 and 90% with fosfomycin based on our observations. To reject the inferiority of fosfomycin with a margin of −7% for CMC, 80% power and 1-sided α of 5%, 188 patients (94 patients per group) would need to be recruited. The selection of −7% as noninferiority margin was decided considering the −10% suggested by the European Medicines Agency for cUTI 18 and given that this study included only bacteremic episodes. For the exploratory study on rectal colonization, a population of 40 patients was targeted.
The differences in proportions with 1-sided 95% CIs were calculated for categorical end points using the comparator group as reference. For secondary outcomes, analyses followed a similar approach. The treatment effect on the primary end point was also analyzed in different subgroups.
Additionally, a multivariable analysis using logistic regression was performed to estimate the impact of treatment on the primary end point, including sites as random effects and other covariates showing a univariate 2-sided P < .20. Significance was set at P < .05 for comparisons not evaluating noninferiority criteria per 95% CIs. Variables not improving the model fit as assessed by using Akaike information criteria were excluded using a stepwise method. For direct comparisons between study groups, 1-sided P values were used. Data were analyzed using SPSS Statistics version 26 (IBM Corp) and R version 3.6.0 (R Project for Statistical Computing) in in May 2021.

Recruitment and Patient Characteristics
Overall, 1578 patients with bacteremia due to E coli were screened; 161 patients were randomized, but 12 patients were found to have exclusion criteria after randomization, 5 patients withdrew consent, and 1 patient was withdrawn by the treating physician. Therefore, 143 patients composed the MITT population ( Figure); 70 patients were assigned to the fosfomycin group and 73 patients to the comparator group (31 patients to ceftriaxone and 42 patients to meropenem). The CEP and MEP comprised 132 and 127 patients, respectively. Completing the recruitment was considered futile, and no additional funding was sought (see subsequent sections). Two hospitals recruited more than 20     to ϱ percentage points; P = .15) ( Table 4). Regarding clinical or microbiological failure, fosfomycin had decreased rates in all subgroups (eTable 5 in Supplement 2). Outcomes were also analyzed among patients who switched to oral drugs (or parenteral ertapenem in the comparator group).
Multivariate analysis was performed to assess the effect of treatment group on CMC, controlling for residual imbalances in exposures. The nonadjusted odds ratio (OR) for CMC among patients receiving fosfomycin, vs patients receiving comparators, was 0.61 (95% CI, 0.28-1.29; P = .20), and In the fosfomycin group, 6 patients (8.6%) developed heart failure (1 patient had 2 episodes, with the second episode occurring after the drug had been discontinued); all these patients were aged 81 years or older, 2 had chronic heart failure, and 3 had chronic kidney insufficiency. Among 5 of these patients, heart failure was considered serious, and among those 5 patients, the drug was discontinued among 4.

Microbiological Studies
Considering

Discussion
In this randomized clinical trial, fosfomycin did not reach the noninferiority criteria in the treatment of bUTI due to MDR E coli. However, this was not due to lack of efficacy; in fact, the clinical and microbiological failure rate was numerically lower with fosfomycin in the MITT, for which the 1-sided 95% CI of the difference was the below the −7% noninferiority margin. in that study were younger and more frequently women than in our study.
Fosfomycin was discontinued among 6 patients because of AEs in our study. This was not the case in the previously mentioned double-blind trial 8 using a similar total daily dose, suggesting a negative impact of the open design against fosfomycin. Nevertheless, heart failure was reported among 6 patients treated with fosfomycin; all but 1 had chronic heart failure (NYHA class I or II) or kidney insufficiency, and all were older than age 80 years. This AE was not described in the cUTI trial, 8 which might be because of the participants' difference in age, and was described among 2 of 2672 patients in a meta-analysis. 6 Heart insufficiency may be caused by the sodium content (14.4 mEq/g) of the intravenous formulation. We suggest avoiding intravenous fosfomycin among patients aged older than 80 years and those with chronic heart or kidney insufficiency. Hypokalemia, usually mild, is a well-known AE associated with fosfomycin. 6,8 The fosfomycin dose used in our study was chosen based on pharmacodynamic data, 24 allowing a 90% probability of target attainment for bactericidal effect for an MIC of 32 mg/L or less. 15 However, selection of resistant subpopulations is a concern; whether it can be avoided by using other dosing regimens is unclear. 23,25,26 In our study, clinical failure due to development of fosfomycinresistant E coli during treatment did not occur.
To the best of our knowledge, this is the first trial to include fosfomycin trometamol as an oral switch among patients with bacteremic infections; its concentrations are high in urine but low in plasma. However, bacteremia in bUTI is an epiphenomenon, and once the parenchymal component of the infection is controlled, urine concentrations may be more important. The outcome data from the subgroup analyses among patients who were switched were encouraging. The investigation of the ecologic impact of the study drugs was exploratory. Overall, the data obtained support the idea that fosfomycin may cause less ecological damage than ceftriaxone and meropenem, and the findings may open the door to further studies.

Limitations and Strengths
This study has several limitations. The calculated sample size was not reached. Additionally, a highly exigent noninferiority margin was chosen. Despite end points being checked by blinded investigators, a lack of blinding may have influenced the delay of hospital discharge and withdrawal of some patients treated with fosfomycin. The options for switching were diverse in the comparator group to mimic standard practice, because the susceptibility of the isolates is unpredictable; however, their efficacy is similar. The rectal colonization study was performed among a small subset of patients.
Some strengths beyond randomization include the pragmatic design, monitoring of quality of data, recruitment of older patients with comorbidities, and exclusion of patients stable enough to