Short Oral Antibiotic Therapy for Pediatric Febrile Urinary Tract Infections: A Randomized Trial Free

This noninferiority, investigator-initiated multicenter, parallel-group, randomized, controlled trial was conducted from May 2020 through September 2022 in 8 pediatric emergency departments in Italy. The trial was conducted following the ethical principles of the Declaration of Helsinki and Good Clinical Practice guidelines and approved by the institutional review board or ethics committee at each center (RC 10-18 prot EudraCT: 2019-001941-41). All parents provided written informed consent.

Inclusion criteria were children 3 months to 5 years of age with a fUTI, no signs of complicated fUTIs (ie, dehydration, vomiting, sepsis), and written informed consent obtained before any study-related assessment. Exclusion criteria were complicated fUTIs (ie, fever persisting more than 48 hours after commencing treatment and/or need to change the antibiotic, dehydration, vomiting, or adherence concern), presence of a urinary catheter, immunodeficiency, hypersensitivity to the antibiotic, antibiotic treatment during the previous 15 days, neurogenic bladder, phenylketonuria or glucose-galactose malabsorption or other underlying metabolic conditions, and intestinal malabsorption.

A fUTI was defined as the presence of fever above 38 °C with a positive urine stick, for nitrite and/or leukocyte esterase, in urine sample collected by clean catch or urinary bag. If the stick resulted positive, subsequent confirmation with a positive urine culture for a single germ obtained by clean catch (≥100 000 CFU/mL) or bladder catheterization (≥10 000 CFU/mL) was performed.⁴ 

According to our national pattern of resistance and updated Italian guidelines, amoxicillin‐clavulanic acid is the first-line oral antibiotic in a well-appearing child with fUTI.⁴  Therefore, patients were randomly assigned to receive amoxicillin-clavulanic acid 50 + 7.12 mg/kg/day orally in 3 divided doses for a total of 5 days (short arm) or 10 days (standard arm). The randomization occurred on the fourth day after enrollment, once the urine culture result was available. Besides age and sex, the type of germ was 1 of the stratification criteria used to ensure an identical distribution of Escherichia coli and non-E. coli in the 2 groups. All patients received a prescription for 5 days of antibiotic treatment, and those randomized to receive 10 days had a second prescription after the randomization (Fig 1). REDCap (v. 8.5.21), a secure web application for building and managing online databases, was used. Patients nonresponsive to treatment (persistence of fever beyond 48 hours) received rescue therapy, namely an alternative antibiotic according to the culture sensitivities, and exited the study. The dosage and administration schedule are those recommended by the Guidelines of the Italian Society of Pediatric Nephrology.⁶ 

The treatment duration was unblinded. Adherence was assessed by verifying the amount of antibiotic remaining in the bottle or the blisters at the reevaluation, 5 days following the therapy’s cessation. Oral paracetamol (15 mg/kg) was prescribed for fever, pain, and irritability. In the event of suspected recurrence (new onset of fever and/or clinical signs and symptoms suggestive of fUTI), the patient was reevaluated whenever necessary, and urinalysis and culture were performed.

The primary end point was UTI recurrence within 30 days following the successful completion of therapy. Recurrences were always symptomatic, and they were considered febrile if the temperature was ≥38 °C, but not if <38 °C.

The secondary end points were:

• clinical recovery at the end of treatment, defined as complete resolution of fever and other signs and symptoms such as irritability, and decreased food intake, if present at onset;

• adverse events related to amoxicillin-clavulanic acid therapy;

• amoxicillin-clavulanic acid and/or other antibiotics resistance of the recurrence assessed through antibiogram.

All children were reevaluated 5 and 30 days after cessation of antibiotic therapy or whenever they had an exacerbation or worsening of symptoms. A urine stick and microscopic examination of the sediment were done at each follow-up to assess possible recurrences: if positive, a urine culture obtained by clean catch or bladder catheterization was performed, and another antibiotic treatment was started.

According to the Italian guidelines, all children underwent a renal and bladder ultrasound after a first fUTI. A voiding cystourethrography was performed only when the ultrasound revealed an abnormality or when the UTI was caused by a pathogen other than E. coli.

To confirm a significant noninferiority of the short antibiotic therapy using a clinically acceptable noninferiority threshold of 5%, we estimated that a sample size of 520 patients (260 per arm) randomly assigned was needed to ensure sufficient discriminatory power. The enrollment of 260 subjects per group would give the study an 80% power to demonstrate the noninferiority of the 5-day treatment compared with the standard one, with an α level of .05. To confirm the noninferiority of the 5-day treatment, the upper limit of the 95% confidence interval (CI) of the difference in recurrence between the 2 groups must remain below the defined noninferiority margin. We also assumed that the number of people lost to follow-up or withdrawn would be at most 10%.

An interim analysis was planned by study protocol 2 years after the first patient was enrolled, by means of the criteria defined by O’Brien-Fleming (presence at the interim analysis of a difference between the two groups with P < .005 [CI at 99.5%]), which allows early closure of the study. The analysis was performed on 142 participants with complete follow-up, 72 in the short arm and 70 in the standard 10-day arm. Because the criteria for early interruption for efficacy were met (see the paragraph on the primary end point in the results), the Steering Committee agreed to stop the trial.

Categorical variables are presented as absolute numbers and percentages. Differences between the 2 groups for categorical variables are assessed with the χ² test or Fisher exact test when appropriate. Continuous variables are presented as means and SDs or medians and interquartile ranges, depending on the data distribution assessed visually and with the Kolmogorov-Smirnov test. Differences between the 2 groups for continuous variables are assessed by Student t test or Mann-Whitney nonparametric test.

Regarding the primary outcome, the principal analysis was carried out both by intention-to-treat (ITT) and per protocol (PP) because CONSORT guidelines strongly recommend providing the details of both estimates in clinical trials: when “ITT and PP provide identical conclusions, the confidence level of the investigator for the study results is augmented.”⁸  For the definition of noninferiority, the difference in recurrence between the 2 groups is calculated with the relative 95% CI. Noninferiority is confirmed if the difference's upper margin of the CI is below the defined noninferiority threshold (5%). As stated previously, fulfilling the criteria defined by O’Brien-Fleming allows early closure of the study. Otherwise, significant differences between the 2 groups with P < .048 will be considered at the end of the study.⁹ 

Secondary outcomes were assessed by comparing categorical variables, presented as absolute numbers and percentages, with the χ² test or Fisher exact test, when appropriate. For secondary outcomes, all children who received the protocol treatment of 5 or 10 days and continued the follow-up were considered.

From May 2020 to September 2022, 175 children were assessed for eligibility, and 142 underwent randomization. The number of patients enrolled by each center is shown in Table 1. Of the 33 excluded, 25 were eligible, but parents declined to participate, 7 had a negative urine culture, and 1 failed to respond to antibiotic treatment within 48 hours (Fig 2). Therefore, a total of 72 patients were randomized to receive amoxicillin-clavulanic acid at a dosage of 50 + 7.12 mg/kg/day orally in 3 doses for a total of 5 consecutive days (short group), whereas 70 patients received the treatment of a total of 10 consecutive days (standard group). Both groups had similar demographic and clinical characteristics at baseline (Table 2). All 142 patients completed the 30-day follow-up.

The prevalent microorganism was E. coli in both groups (86.1% in the short and 88.6% in the standard group; P = .66).

The 2 patient arms had a similar incidence of recurrent UTIs before the study (19.4% vs. 11.4%, P = .19), as well as a similar presence of anatomic risk factors like vesicoureteral reflux (VUR) and urinary tract dilation¹⁰  already known before enrollment (6.9% vs. 7.1%, P = 1.0).

ITT and PP analyses provided identical conclusions because all participants fully adhered to the follow-up required by the study protocol. Particularly, at the 2-year interim analysis, the recurrence rate of all UTIs (febrile and nonfebrile) within 30 days from the end of therapy was 2.8% (2/72) in the short group and 14.3% (10/70) in the standard group (Table 3); thus, the difference between the 2 groups was –11.51% (95% CI, –20.52 to –2.47). This result supports the noninferiority of a 5-day treatment because the upper CI falls below the defined threshold of +5%. The following values were obtained by increasing the CI from 95% to 99.5% (corresponding to P < .005): difference –4.33%; 99.5% CI, –13.02% to 4.37%, thus confirming the noninferiority of short therapy and the ability to discontinue the study.

Resolution of signs and symptoms (clinical well-being) occurred in 97.2% (70/72) of cases in the short group and 92.9% (65/70) of cases in the standard group (P = .27) (Table 4). The need for further antibiotic therapy was 1.4% (1/72) in the short group and 5.7% (4/70) in the standard group (P = .21). Antibiotic treatment was conversely administered to 1/72 (1.4%) in the short group and 6/70 (8.6%) in the standard group (P = .06) at the second follow-up after 30 days. The frequency of adverse events was 1.4% (1/72, diarrhea) in the short group and none in the standard group (P = 1.0).

The emergence of antibiotic resistance to amoxicillin-clavulanic acid or opportunistic pathogens after the first-line therapy did not statistically differ between the two treatment groups in the first and second follow-ups. Supplemental Table 6 shows the isolated uropathogen in the first and second urine cultures between the patients who experienced a recurrence.

No difference was present as in terms of urinary tract abnormalities and VUR between the 2 groups (Table 5): some patients had a known VUR at randomization (5 in the short arm and 2 in the standard one), whereas in other children, reflux was diagnosed at the cystourethrography performed when suggested by our Italian guidelines, following the urinary tract infection (4 in the short arm and 8 in the standard one).

Because the presence of VUR and non-E. coli UTIs could be associated with a higher risk of recurrence, we performed 2 post hoc analyses excluding patients with known VUR at the enrollment or patients with non-E. coli UTIs. Both analyses confirmed the noninferiority of short therapy, also increasing the CI from 95% to 99.5% (Supplemental Tables 7 and 8).

In this study, a 5-day treatment with oral amoxicillin-clavulanic acid of an acute episode of fUTI was not inferior to a 10-day course, in terms of clinical cure rates, recurrence of infection within 30 days from the end of therapy, adverse events, and development of antibiotic resistance.

Published guidelines have shown a lack of consensus regarding the duration of antibiotic therapy for fUTIs; this is often based on anecdotal data, clinical intuition, habit, or “expert opinion.”¹¹  In this context, a multicenter, randomized clinical trial¹²  comparing oral versus initial parenteral antibiotic therapy showed that urine culture was negative within 24 hours in all children in both groups. Furthermore, the mean time to defervescence was 25 and 24 hours for children treated orally and intravenously, respectively. The same data emerged in another study conducted on 502 patients, in which the time to defervescence was less than 48 hours.¹³ 

Studies regarding other bacterial infections in children have been carried out to investigate the effectiveness of shorter antibiotic treatments. In streptococcal pharyngitis, 3 to 6 days of antibiotics (amoxicillin, macrolides, cephalosporins) were equivalent to 10 days of penicillin V.¹⁴  Furthermore, few patients completed the 10-day course of penicillin V, making this treatment regimen of questionable value. In pediatric community-acquired pneumonia, the SAFER study demonstrated that 5 days of amoxicillin therapy (followed by 5 days of placebo) was comparable to a 10-day course.¹⁵  The SCOUT-CAP study¹⁶  compared 5 and 10 days of oral therapy with β-lactams for the outpatient treatment of pediatric community-acquired pneumonia through a multicenter randomized clinical trial. Of the children who responded to the initial treatment, the 5-day therapy was superior to the 10-day therapy, with no significant differences in clinical response and the incidence of adverse effects to antibiotics.

To our knowledge, this is the first study addressing the noninferiority of a shorter treatment in fUTI, except for a single abstract dealing with a series of children aged 2 to 10 years with both lower and upper UTIs, presented at the ID International meeting (IDSA 2020).¹⁷  Ten days of antibiotic treatment is recommended by our national guidelines⁴  and therefore was the length of the treatment of the standard arm. A 5-day course was chosen for the short arm, based on similar studies on pneumonia in children previously mentioned.^15,16 

Our study also enrolled children with a history of recurrent UTIs (Table 2). In the recurred patients, there was only 1 with a positive history for infection recurrence, randomized in the standard arm, even though the number of patients with recurrent UTIs was slightly higher in the short treatment group (19.4% vs 11.4%, P = .19). Because there were no significant differences between the 2 groups, we can assume that short-term therapy may also be adequate for children with recurring UTIs. The incidence of anatomic abnormalities and VUR was the same in the 2 groups, ruling out the hypothesis that the higher number of recurrences of UTI in the standard treatment group could be due to a higher prevalence of underlying abnormalities.

The choice to separately analyze the recurrence of fUTIs was motivated by the fact that fever is suggestive of kidney parenchymal involvement. In nonfebrile UTIs, it is often difficult to establish the site of the infection because fever may be absent in infants aged 3 months or younger.^1–3  Consequently, because children aged 3 months or older who presented symptoms of the lower urinary tract were included among the recurrences, it was decided to carry out a diversified analysis between nonfebrile and fUTIs. The consequent analysis confirmed the noninferiority of short therapy.

In this study, the occurrence of antibiotic resistance or opportunistic pathogens of the recurrences, such as the onset of UTI with an amoxicillin-clavulanic acid–resistant bacterium, did not differ between the 2 groups, although the small sample size does not allow us to generalize the result. This result argues against the assumption commonly believed that antibiotic therapy should be continued for days after complete remission of febrile infection to avoid antibiotic resistance, which has been handed down for years without objective evidence.

The interim analysis was conducted as planned in the protocol 2 years after the first enrollment. The study started in 2020, and there was still uncertainty regarding the evolution of the COVID-19 pandemic. During the 2 years of study, there was a low recruitment rate because of the reduced access to the pediatric emergency departments.^18,19  The interim analysis showed a recurrence rate of all UTIs (febrile and nonfebrile) within 30 days from the end of therapy of 2.8% (2/72) in the short group and 14.3% (10/70) in the standard group. This difference in favor of the short group was unexpected. Because the use of the O’Brien-Fleming criteria, with the increase of the CI from 95% to 99.5%, confirmed the noninferiority results, the Steering Committee decided that the results were robust and agreed to stop the trial, in the light of a low recruitment rate and to allow us to highlight the short therapy’s effectiveness promptly.

A limitation of the study is that the randomization of therapy duration was unblinded. All parents were aware of the ongoing trial and may have been sensitized to look for possible infection symptoms (anorexia, irritability, urine smell), taking their kids for an evaluation before the appearance of high fever. No recurrence was found at the 5- and 30-day scheduled controls, but all relapsing children were referred for symptoms by the parents. Further larger studies are needed to better address this issue. Another limit is represented by the fact that drug adverse effects were measured by asking the patient’s parents at the reevaluation 5 and 30 days after cessation of antibiotic therapy and no standardized questionnaire was used.

The strengths of this study are the prospective multicenter design, the homogeneity of the enrolled population in terms of age, which includes mostly infants and toddlers, the analysis of fUTIs among recurrences, and the thorough imaging evaluation received by all patients according to the guidelines.

To our knowledge, this is the first study in the literature concerning brief therapy in UTIs. Further trials are needed to confirm these results.

In this randomized controlled study, a 5-day treatment of acute fUTIs in children with oral amoxicillin-clavulanic acid was noninferior to a 10-day treatment. Further data are needed to confirm our results.

The authors thank Giulia Pennesi for the English revision of the manuscript.

STOP Trial Group members: Stefano Amoroso, MD (Department of Pediatrics, “S. Maria degli Angeli” Hospital, Pordenone); Roberto Chimenz, MD (Department of Pediatrics, “Gaetano Martino di Messina” Hospital, Messina); Ester Conversano, MD (Institute for Maternal and Child Health-IRCCS 'Burlo Garofolo', Trieste); Antonio Gatto, MD (Department of Pediatrics, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome); Federico Marchetti, MD (Department of Pediatrics, “S. Maria delle Croci' Hospital, AUSL della Romagna, Ravenna); Cristina Milocco, MD (Department of Pediatrics, “S. Polo” Hospital, Monfalcone); Giuliana Morabito, MD (Department of Pediatrics, “S. Maria degli Angeli” Hospital, Pordenone); Floriana Scozzola, MD (Pediatric Unit, Ca' Foncello's Hospital, 31100 Treviso); and Federica Vianello, MD (Pediatric Nephrology, Dialysis and Transplant Unit, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, via della Commenda 9, Milano), all in Italy.

CI

confidence interval

fUTI

febrile urinary tract infection

ITT

intention to treat

PP

per protocol

UTI

urinary tract infection

VUR

vesicoureteral reflux