Short Intravenous Antibiotic Courses for Urinary Infections in Young Infants: A Systematic Review Free

The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline²¹  was followed in this review, and the protocol was registered in PROSPERO (ID: CRD42021215114). On February 1, 2021, 4 databases (PubMed, the Cochrane Library, Medline [Ovid] [starting from 1946], and Embase [Ovid] [starting from 1947]) were searched with assistance from an expert librarian using the following search terms: urinary tract infections, urinary-pathogen*, UTI, urinary infection*, pyelonephritis, cystitis, bladder-inflammation, anti-bacterial agents, anti-infective agent, antibiotic, antimicrobial*, ampicillin, amoxicillin, amoxicillin-clavulanate, cephalosporin*, cefotaxime, ceftriaxone, cefixime, cephalexin, ceftibuten, cefaclor, aminoglycoside*, gentamicin, amikacin, Escherichia coli, e-coli, Enterococcus faecalis, e-faecalis, Staphylococcus, Staphylococcus-saprophyticus, s-saprophyticus, Klebsiella, Enterobacter, Enterococcus, Proteus-mirabilis, p-mirabilis, Pseudomonas-aeruginosa, p-aeruginosa, bacteremia, newborn*, baby, babies, neonat*, infan*, 1-month, 2-month*, 3-month*, one-month, two-month*, three-month*, 30-day*, 60-day*, 90-day*. The complete search strategy is outlined in Supplemental Table 6. Results were limited to English language and human studies.

Titles and abstracts of articles retrieved from the search strategy were independently screened for eligibility by 2 reviewers (S.H. and J.L.). Following this, full texts of potentially relevant papers were obtained and assessed for final inclusion by 1 reviewer (S.H.). Disagreements over the eligibility of studies were resolved by consensus and, if necessary, through discussion with a third reviewer (A.G.). The reference lists of relevant articles were also manually checked to identify additional studies.

Articles meeting the following criteria were included: original study; those involving infants aged ≤90 days with a UTI; those that specified the duration of IV antibiotic treatment; and those that reported 1 or more treatment outcomes. Only studies involving shorter IV antibiotic courses (mean and median of ≤7 days) or those involving only oral therapy were included. Authors were contacted to request further information where specific data for infants aged ≤90 days were not reported. Conference proceedings, meeting abstracts, and case reports with <10 young infants were excluded. For all studies meeting the inclusion criteria, data pertaining to study design, participants, clinical presentation (concomitant fever, bacteremia, meningitis, renal tract anomalies), antibiotic route and duration, and clinical outcomes were extracted.

Risk of bias was assessed by 1 reviewer (S.H.) by using the Newcastle-Ottawa Scale (NOS)²²  for observational studies and the Revised Cochrane Risk-of-Bias Tool for randomized controlled trials (RoB 2).²³  The Oxford Centre for Evidence-Based Medicine (CEBM) Levels of Evidence²⁴  was used to grade the strength of the proposed clinical recommendations for managing UTIs in young infants.

The search yielded 6665 articles in Embase, 4336 in Medline, 2364 in the Cochrane Library, and 792 in PubMed. After removing duplicates, the remaining 10 181 records were screened, and 18 articles that met the inclusion criteria were identified (Fig 1).^{12,15–17,25–38}  Two were randomized controlled trials (RCTs),^27,32  and 16 were retrospective studies,^* which overall included 16 615 infants aged ≤90 days. The key results of each study are summarized in Tables 1–4, and the proposed clinical recommendations are outlined in Table 5.

Bias assessment for the studies is presented in Supplemental Tables 7 and 8. For the 2 RCTs, the overall risk of bias was rated as “some concerns” using the RoB 2 tool.^27,32  Of the 16 observational studies, 3 had a low risk of bias,^12,15,29  and 13 had a moderate to high risk of bias using the NOS tool.^†

Definitions of UTI included a positive urine culture in all studies.^{12,15–17,25–38}  However, bacterial counts varied depending on the sampling method (Supplemental Table 9). Five studies did not report the urine collection method or define the required bacterial colony count for UTI.^{12,15,17,27,38}  In 1 study, infants with negative urine cultures were also included if the urine sample was obtained while receiving antibiotics and if the urinalysis was suggestive of UTI with the presence of leukocyte esterase, nitrites, and a high leukocyte count.²⁵ 

Initial IV antibiotics were used for the treatment of all young infants with UTIs in 15 of the 18 studies.^‡ In the remaining 3 studies, which included only infants aged ≥30 days, oral antibiotics alone were used for all infants (1 study³⁵ ) or a proportion of infants (2 studies^27,32 ).

Thirteen studies included infants with bacteremic UTI; however, only 6 reported their specific treatment duration and outcomes.^{16,17,29,31,33,38}  This included a total of 468 infants aged ≤90 days with bacteremic UTI.

The largest 2 studies were multicenter retrospective cohort studies including only young infants with bacteremic UTI.^17,29  Desai et al²⁹  studied 115 young infants aged ≤60 days comparing those who received a ≤7 days to >7 days IV antibiotic course (having excluded meningitis) and found no significant difference in the 30-day UTI recurrence (3% vs 7%; adjusted relative risk [aRR] 1.9; 95% confidence interval [CI], 0.3–11.6) or 30-day all-cause hospitalization (10% vs 16%; aRR 1.2; 95% CI, 0.4–3.9). Similarly, in a study of 251 young infants aged <90 days, 57% of whom received ≤7 days of IV antibiotics, Schroeder et al¹⁷  found the unadjusted mean IV antibiotic duration did not differ significantly between those with and without 30-day UTI recurrence (8.2 days vs 7.8 days; P = .81). Notably, in this study meningitis was excluded in only 196 (78%) infants, and the authors did not assess all-cause hospitalization after treatment.¹⁷  In both studies, infants with prolonged bacteremia, defined as a positive blood culture for >1 day, received longer IV antibiotic courses (additional 3.5 days¹⁷  or >7 days²⁹ ). Overall, there were no severe UTI-related complications (admission to intensive care, requirement for mechanical ventilation or inotropes) in either study, and the 30-day UTI recurrence was low (5%²⁹  and 2%¹⁷ ) with the majority (67% to 100%) of infants with recurrence having vesicoureteral reflux (VUR).^17,29  A limitation of both studies is that neither accounted for cases of UTI recurrence presenting to other health centers.^17,29 

The remaining 4 retrospective studies included infants with both bacteremic and nonbacteremic UTI, with a total of 102 young infants having bacteremic UTI.^16,31,33,38  Although none of the included infants had concurrent meningitis, a lumbar puncture was not routinely performed in 2 studies,^16,33  whereas another study³¹  did not report whether meningitis was excluded. The median duration of IV antibiotics ranged from 4 to 7 days.^16,31,33,38  In 3 studies,^16,31,33  there were no episodes of UTI recurrence within 14 to 30 days, whereas, in the fourth study,³⁸  the authors reported a 10% recurrence rate (3 of 29 infants). In the latter study, the median IV antibiotic duration for bacteremic UTI was 4.9 days (IQR 3.3–7); however, the exact treatment duration for those with recurrence was not reported.³⁸  Notably, 2 of the 3 infants with recurrence had renal tract anomalies (Sheila Swartz, MD, MPH, written communication, November 30, 2020).³⁸  In the only study in which the authors reported the incidence of renal scarring, it was found that 2 of 16 (12.5%) neonates with bacteremic UTI had scarring at 6 months.³³ However, results for this outcome were only available for three-quarters of participants.³³  In this study, the mean IV antibiotic duration for bacteremic UTI was 7 days; however, the length of IV treatment of the 2 neonates with scarring was not reported by the authors.³³ 

Twelve studies included 15 826 young infants with nonbacteremic UTI.^§ Two were large multicenter retrospective cohort studies that compared IV antibiotic treatment of ≤3 days to >3 days and found no significant difference in the adjusted 30-day UTI recurrence between the 2 groups.^12,15  Brady et al¹²  studied 12 333 infants aged <6 months (0.7% with bacteremia) of whom the majority (74%) were <3 months old but did not report specific outcomes for the young infant cohort. However, in a subanalysis of 1-month age increments, the authors found no association between the IV antibiotic duration and UTI recurrence.¹²  In addition to UTI recurrence, Lewis-de Los Angeles et al¹⁵  also reported the 30-day all-cause hospitalization in 3973 infants aged ≤60 days in whom meningitis and bacteremia were excluded, and they found no significant difference between short (≤3 days) and long (>3 days) IV antibiotic treatment (4.1% vs 5.0%; adjusted odds ratio [aOR] 1.16; 95% CI, 0.83–1.62).

Rates of UTI recurrence in young infants with nonbacteremic UTI were also assessed in 4 smaller retrospective studies and were found to be low (0% to 1.8%).^31,33,34,38  The median duration of IV antibiotic treatment in these studies ranged from 2 to 4 days with authors of 2 studies also reporting low rates of 30-day all-cause hospitalization (3.7%³⁸  and 5.4%³⁴ ) with a median of 2 days of IV antibiotics.

Complications of UTI were reported by the authors of 4 studies and included renal scarring (1 study³³ ) and severe illness (3 studies^26,28,36 ). A study in neonates found that 9 of 114 (7.9%) had renal scarring at 6 months when treated with a median IV antibiotic duration of 4 days.³³  Severe illness associated with UTIs was infrequent (0% to 7%) in the 3 retrospective studies in which the authors reported this outcome.^26,28,36  Schnadower et al³⁶  found that 41 of 1754 (2.3%) infants aged 29 to 60 days who were treated with a median IV antibiotic duration of 4 days (IQR 3–5) experienced 1 or more of the following adverse events: admission to the ICU, death, shock, bacterial meningitis, need for ventilatory support, need for surgery, or other clinical complications (including seizures, coagulopathy, and congestive heart failure). Similarly, a study of 67 young infants treated with a median IV antibiotic duration of 7 days found that 5 (7%) required ICU admission, although none developed septic shock or died.²⁶  The final study reported no ICU admissions in a cohort of 123 young infants treated with a median of 3.5 days of IV antibiotic therapy.²⁸ 

Four retrospective studies reported the treatment outcomes of infants with concurrent sterile CSF pleocytosis treated with shorter (≤7 days) IV antibiotic courses.^25,33,34,37 

The largest study was a multicenter retrospective study that compared 214 infants with CSF pleocytosis (8% with bacteremia) with 976 infants with normal CSF (6.2% with bacteremia) who were treated with IV antibiotics for a median of 4 days (IQR 4–6) and 3 days (IQR 3–5), respectively.³⁷  The authors of this study found no significant difference in the rate of adverse events provided that the young infant was “not clinically ill” (defined by the absence of dehydration, respiratory distress, concomitant infections) and did not have a significant past medical history.³⁷  Notably, this study did not include neonates.³⁷  The authors of the remaining 3 studies reported favorable outcomes (clinical improvement, no UTI recurrence) in 20 young infants with nonbacteremic sterile CSF pleocytosis treated for a mean IV duration of 2 to 5 days.^25,33,34 

Oral antibiotic treatment alone was reported in 3 studies,^27,32,35  including 2 RCTs^27,32  of infants with febrile UTI where meningitis was excluded. Hoberman et al³²  compared oral cefixime alone to initial IV cefotaxime for 3 days or until the infant was afebrile for 1 day (whichever was longer), followed by oral cefixime; they found no significant difference in UTI recurrence and renal scarring at 6 months. Interestingly, all urine cultures (obtained from 95% of included children) were sterile within 24 hours of treatment.³²  This trial included 306 children aged 1 to 24 months of whom 4% were bacteremic; however, it is unclear how many of these infants were ≤90 days old.³²  Similarly, Bocquet et al²⁷  randomized 171 children with UTIs of whom 21 (12%) were 1 to 3 months old (10 received oral cefixime alone, and 11 received IV ceftriaxone for 4 days followed by oral cefixime). The frequency of renal scarring at 6 to 8 months for these infants was 0 of 4 (0%) in the oral group and 2 of 6 (33%) in the IV followed by the oral group.²⁷  However, this outcome was only assessed for those with abnormal dimercaptosuccinic acid scans during the acute period (4 of 10 in the oral group and 6 of 11 in the IV group).²⁷  In this study, it was also unclear whether bacteremia was excluded.²⁷ 

The only study of oral antibiotics alone for UTIs in young infants was a single center retrospective study by Pennesi et al.³⁵  Of 51 infants aged 1 to 3 months who received 1 of the following oral antibiotics (ceftibuten, cefaclor, or amoxicillin/clavulanate) for 10 days, none had UTI recurrence.³⁵  Only the 4 infants with VUR had a dimercaptosuccinic acid scan at 6 months and 3 of 4 had renal scarring, all of whom had grade IV VUR (Marco Pennesi, MD, written communication, September 2, 2020).³⁵  In this study, the authors did not report the presence of concurrent bacteremia or meningitis.³⁵ 

This is the first systematic review of the evidence for the optimal timing of IV to oral antibiotic switch for UTIs specifically in infants aged ≤90 days. Our review found that a shorter duration of IV antibiotic treatment of ≤7 days for bacteremic UTI and ≤3 days for nonbacteremic UTI is not associated with an increased rate of UTI recurrence, all-cause hospitalization, or adverse clinical outcomes. Notably, these recommendations are only for young infants where meningitis has been excluded and who have no features of sepsis. These findings have important implications for clinical practice because shorter IV antibiotic courses with an early switch to oral therapy can improve the quality of life of children and their families, reduce the length of hospital stay, the risk of nosocomial infections, and health care costs.^19,20 

We also found that oral antibiotics alone can be considered for infants aged 1 to 3 months, although further high-quality studies specifically in this age group are needed. This aligns with the current management approach for children aged >90 days where oral antibiotics are the treatment of choice for UTIs, unless the child is severely unwell or unable to take oral antibiotics.^11,18,39,40  Concerns about oral therapy for UTIs in infants aged ≤90 days are, firstly, the potentially lower bioavailability of oral antibiotics in young infants⁴¹  and, secondly, the higher risk of complications from pyelonephritis.^5,18  A systematic review of 31 studies in neonates revealed that although the bioavailability of oral antibiotics is lower than that of parenteral antibiotics, adequate serum concentrations for bacterial killing are achieved in most neonates.⁴¹  Furthermore, commonly used oral antibiotics for UTI such as co-trimoxazole and cefalexin have been shown to achieve high concentrations in urine.^42–44  In young infants, the frequency of concurrent pyelonephritis is reportedly high (60% to 65% in febrile UTI),^45,46  and it is difficult to clinically distinguish between pyelonephritis and cystitis because infants often lack the classic symptoms.^3–5  However, in this review, the presence or absence of pyelonephritis was only reported by authors in 8 studies and did not appear to significantly influence the IV antibiotic duration.^{12,15,25,27,28,30,32,33} 

Although sterile CSF pleocytosis in the setting of UTIs has been reported in 12.8% and 23% of infants aged ≤90 days,^47,48  its etiology and implications for antibiotic treatment are unclear. Some studies hypothesized that the systemic inflammatory cytokines released in response to uropathogens might play a role in the pathophysiology of CSF pleocytosis through triggering innate immune system activation in the brain.^37,47–50  Our review identified 1 large retrospective study which found that a median of 4 days of IV antibiotics in well-appearing infants with uncomplicated medical histories was not associated with adverse events.³⁷  However, to date, no studies have evaluated the long-term neurodevelopmental outcomes of this patient population, and further prospective studies are needed.

The most common clinical outcome evaluated in the included studies was the frequency of UTI recurrence, which overall was low (0% to 7%).^{12,15–17,25–38}  The authors of 3 of the 6 studies that included young infants with renal tract abnormalities found that 60% to 100% of UTI recurrences occurred in those with anomalies.^17,29,38  This supports existing literature showing that underlying renal tract abnormalities, particularly VUR, increase the risk of recurrent UTIs.^51–53  Similarly, high-grade VUR is a known risk factor for renal scarring with 68.6% developing this complication.⁵⁴  Therefore, for infants in this review who developed renal scarring or UTI recurrence in the presence of renal tract abnormalities, it is likely that the underlying anomalies had a greater influence on clinical outcomes rather than the route or duration of IV antibiotic treatment.

Whereas there is some consensus on the approach to UTI treatment in older infants and children, guidelines on the management of UTIs in infants aged ≤90 days are scarce.¹³  The authors of 4 systematic reviews studied the timing of IV to oral antibiotic switch in children with UTIs, but unfortunately, none of them performed a subanalysis for infants aged ≤90 days.^55–58  This lack of evidence reflects the wide variation in practice.^{12,15–17,25–38}  Clearly, further dedicated studies in young infants are required.

Definitions of UTI and UTI recurrence varied between studies and the majority did not specify the rates of pyelonephritis in the young infants. Additionally, information on the use of prophylactic antibiotics, latency between symptom onset and UTI diagnosis, and whether the UTI represented a first or recurrent episode was missing in several studies. Eleven of the 18 studies only reported the duration of IV antibiotics and did not state the total antibiotic duration. Also, none of the studies determined whether infants with UTI recurrence presented to other health care centers, potentially resulting in underestimation of the recurrence rates. Finally, the majority of studies were observational with only 2 RCTs,^27,32  both of which included a limited number of infants aged ≤90 days and had “some concerns” for bias. Importantly, results of this review cannot be generalized to infants with UTI and concurrent meningitis because only 1 such infant²⁸  was identified in this review.

UTIs are the most common serious bacterial infection in infants aged ≤90 days.¹  This review has shown that a shorter IV antibiotic duration of ≤7 days for bacteremic UTI and ≤3 days for nonbacteremic UTI should be considered for infants aged ≤90 days, provided that meningitis is excluded and there are no features of sepsis. The existing data suggest that shorter IV antibiotic courses can be used in well-appearing infants with concurrent sterile CSF pleocytosis and oral antibiotics alone can be considered for UTIs in infants aged 1 to 3 months; however, further studies are needed.

We thank Poh Chua, expert medical librarian at The Royal Children’s Hospital Melbourne, for her assistance with the literature search. She was not compensated beyond her usual salary for her work on this study.

AB

antibiotic

aOR

adjusted odds ratio

aRR

adjusted relative risk

CI

confidence interval

CSF

cerebrospinal fluid

IQR

interquartile range

IM

intramuscular

IV

intravenous

NOS

Newcastle-Ottawa Scale

RCT

randomized controlled trial

RoB 2

Revised Cochrane Risk-of-Bias Tool for randomized controlled trials

UTI

urinary tract infection

VUR

vesicoureteral reflux