Urinary tract infections (UTIs) are among the most frequent bacterial diseases in febrile infants.1–5 In incontinent patients, they are usually detected in emergency departments (EDs) due to the need to collect a sterile urine sample for diagnosis. In addition to the diagnosis and the initial course of antibiotherapy, management should include informing the family on the indications for subsequent imaging tests.

Although there are discrepancies between current guidelines and between clinical protocols at different sites regarding the indications for the various imaging tests that can be performed after a first UTI,6 renal ultrasound is included in most due to its broad availability and lack of harmful effects. However, since the introduction of routine prenatal ultrasound examinations, several publications have argued that the most frequent urinary tract malformations are detected before birth.7 With the aim of omitting performance of renal ultrasound after an initial episode of UTI in patients with a normal prenatal ultrasound, Miron et al.8 conducted a prospective study in 250 children aged less than 5 years and assessed the correlation between prenatal and postnatal ultrasound findings, finding a concordance of 96% (95% confidence interval [CI], 93–99). Nevertheless, the role of prenatal ultrasound is still being debated, and the ultrasound after the first UTI is included in many guidelines. Thus, the clinical practice guideline for the management of UTI of the American Academy of Pediatrics9 recommends performance of a renal ultrasound in any febrile child aged less than 2 years with a first episode of UTI. In Spain, both the UTI protocol of the Asociación Española de Pediatría (Spanish Association of Pediatrics)10 and the clinical practice guideline for the management of UTI in the paediatric population developed by the Ministry of Health11 recommend its performance after an initial febrile episode of UTI, with the latter adding the following criteria: incontinent patient without a normal prenatal or postnatal ultrasound, signs of urinary tract dysfunction, abdominal or vesical mass, elevated serum creatinine levels or UTI caused by an organism other than Escherichia coli.

The objectives of our study were to assess the diagnostic yield of renal ultrasound after an initial UTI in patients aged less than 2 years and to analyse potential risk factors for abnormal ultrasound findings.

A total of 627 episodes of UTI were diagnosed during the period under study, with microbiological confirmation in 480 cases. After excluding 174 patients for meeting exclusion criteria, the final sample comprised 306 cases (Fig. 1). Table 1 describes the clinical and epidemiological characteristics of the sample.

Figure 1.

Flow chart of the study.

DUS, disease of the urinary system; US, ultrasound; UTI, urinary tract infection.

(0.08MB).

The renal ultrasound was abnormal in 35 patients (11.4%; 95% CI, 8.3–15.5): detection of urinary tract dilatation of varying degree in 24 (68.6%) and structural anomalies in 11 (31.4%). The nephrology department decided to expand the evaluation with a voiding cystourethrogram (VCUG) in 26 patients (74.6%), which led to diagnosis of vesicoureteral reflux (VUR) in 7 (26.9%). In 6 cases (2.0%; 95% CI, 0.9–4.2), the renal ultrasound was considered pathological, as it led to prescription of antibiotic prophylaxis. These cases included 4 children with high-grade VUR (iv–v) and 2 with structural anomalies (megaureter). These last 2 patients were also referred for assessment to the Department of Surgery.

In the univariate analysis (Table 2), male sex, age less than 3 months and a causative pathogen other than E. coli were predictors of abnormal ultrasound findings. In the multivariate regression analysis, the variables that remained significant were age less than 3 months, with an OR of 2.1 (1–4.3; P=.05) and a causative pathogen other than E. coli, with an OR of 3.8 (1.7–8.7; P=.002). The 6 patients with a pathological renal ultrasound presented one or both of these risk factors.

Of the 271 patients with a normal renal ultrasound, 26 (9.6%) had a new episode of UTI in the first year of follow-up. A new renal ultrasound was performed in 7 of them, and was abnormal in 1 case (mild ectasia). Fifteen of these patients were evaluated with a VCUG, which detected a low-grade VUR in 3.

Our study confirmed the low yield of renal ultrasound following a first UTI, as the test was only relevant in patients with risk factors. Several recent studies have found similar results concerning the efficacy of renal ultrasound in patients with UTI as a tool for screening for urinary system anomalies, with a low sensitivity, specificity and predictive value compared to other diagnostic techniques (VCUG, dimercaptosuccinic acid [DMSA] renal scan).13,14 In 2014, Zamir et al. published a study in 255 patients with a first UTI; the renal ultrasound was abnormal in 33, and the management was not changed as a result of the findings in any of the patients.15 The results of a study by Nelson et al. were similar, and the authors concluded that renal ultrasound is a poor tool for diagnosis of urinary system anomalies after analysing a sample of 2259 patients and finding a prevalence of clinically significant anomalies of 2%.16 These results agree with those of our study, where the renal ultrasound after a first UTI only led to changes in management in 6 patients (2% of cases), all of them aged less than 3 months and/or with a UTI caused by a pathogen other than E. coli. The latter finding is consistent with the findings of a previous study conducted in the same hospital17 and with the indications of the clinical practice guideline of the Spanish National Health System of 2011,11 which specifically recommends performance of a renal ultrasound in patients with UTI caused by a microorganism other than E. coli on account of the existing evidence on the high prevalence of urinary system abnormalities in these patients. On the other hand, the higher frequency of abnormal findings in the follow-up ultrasound in younger children, which is statistically significant in infants aged less than 3 months, could be partly explained to the higher prevalence of VUR in younger children, which decreases gradually with age (with a rate of spontaneous resolution of 10% to 15% per year of growth).18

Multiple studies have demonstrated that patients with diseases of the urinary system are at higher risk of urinary tract infection; previous calculations19 have estimated the positive predictive value of recurrent UTI for predicting high-grade VUR at 55%, whereas children with a single episode of UTI are at lower risk of associated malformations. Based on this evidence, the latest update of the National Institute for Health and Care Excellence (NICE) guideline on the management of UTIs in patients aged less than 16 years20 recommends against routine performance of ultrasound after a first uncomplicated UTI in children aged more than 6 months, reserving this diagnostic tool for cases of recurrent UTI, atypical UTI or patients aged less than 6 months. The low frequency of recurrence of UTI within a year of diagnosis in patients with a normal ultrasound after a first episode of UTI found in our study supports this recommendation, and suggests that it may be more useful to perform a VCUG after a second episode of UTI rather than a second ultrasound in patients without any of the 2 risk factors mentioned above.

The main limitations of our study are those intrinsic to its retrospective design. On the other hand, while the regionwide electronic health records allowed us to track the outpatient management of the patients, we had to exclude 17 patients due to the absence of records for a prenatal or post-UTI ultrasound; it is possible that these patients were managed in private facilities and that there were radiologic findings unknown to us in these cases. However, this subset of patients is very small, and therefore we consider that it did not have a relevant impact on the results. For the same reason, it is also possible that some of the patients with normal ultrasounds had subsequent episodes of UTI during the follow-up that we were unable to detect.

To conclude, the overall yield of renal ultrasound after a first episode of UTI in children aged less than 2 years is low, as its findings have little impact on the subsequent management in most cases, which suggests that the decision to perform a renal ultrasound should be made on a case-to-case basis based on the presence of risk factors: age less than 3 months or causative microorganism other than E. coli.

The authors have no conflicts of interest to declare.