Abnormal Ultrasound Findings After a Urinary Tract Infection in Children

ABSTRACT

Aim

This study aimed to investigate the occurrence of urinary tract abnormalities detected in the ultrasound screening of children after a urinary tract infection episode, regardless of their age.

Methods

This cohort study included children aged 0–16 years who underwent urinary tract ultrasound screening after a urinary tract infection between January 1, 2004, and December 31, 2015 in a paediatric university hospital in Finland. We excluded children with previously diagnosed pathologies of the urinary tract.

Results

A total of 2050 children were included in this study. Among these, 116 (5.7%, 95% confidence interval 4.7%–6.7%) children had an abnormal ultrasound. In children with pyelonephritis, 70% of the abnormal ultrasounds were found in children younger than 2 years old. Age younger than 2 years was not associated with an increased risk of abnormality (odds ratio: 0.99, 95% confidence interval 0.65–1.52). Boys and children with non‐ Escherichia coli urinary tract infection or recurrent urinary tract infection were associated with a higher risk of having a urinary tract abnormality.

Conclusion

Limiting ultrasound screening after pyelonephritis to children under 2 years of age, as recommended in many current treatment guidelines, may miss a significant proportion (30%) of children with urinary tract abnormalities.

Abbreviations

AAP

The American Academy of Pediatrics

CI

confidence interval

EAU/ESPU

European Society for Paediatric Urology

NICE

The National Institute for Health and Care Excellence

OR

odds ratio

US

ultrasound screening

UTI

urinary tract infection

Summary.

• Most treatment guidelines currently recommend ultrasound screening of the urinary tract for children younger than 2 years of age after a urinary tract infection.

• In this study of 2050 children, we demonstrated that limiting ultrasound screening after pyelonephritis to children < 2 years of age may miss a marked proportion (30%) of children with abnormalities.

• Our results implicate a need to review the age criteria proposed in the current guidelines.

1. Introduction

Ultrasound screening (US) of the urinary tract is commonly recommended for children after a urinary tract infection (UTI) to check for abnormalities, including obstruction, renal structural anomalies, nephrolithiasis, or an abdominal mass [1]. The American Academy of Pediatrics (AAP) recommends US after the first febrile UTI in children aged 2 months to 24 months [2]. The National Institute for Health and Care Excellence (NICE), United Kingdom, recommends US after the first UTI in children younger than 6 months, and for children with atypical UTI and recurrent UTIs [3]. The European Society for Paediatric Urology (EAU/ESPU) recommends US after the first febrile UTI in infants, whereas the Asian guidelines recommend US after the first febrile UTI regardless of age [4, 5].

Notably, the selection criteria recommended for children to undergo US after UTI in the present guidelines [2, 3, 4, 5] are based on studies that have mainly included children younger than 2 years of age [6, 7, 8, 9, 10]. About 12%–27% of children included in these observational studies were found to have urinary tract abnormalities in the US performed after a UTI [6, 7, 8, 9, 10]. Furthermore, one study including children exclusively older than 5 years reported a 17% incidence of urinary tract abnormalities in children undergoing US after a UTI [11]. Evidently, there is a lack of strong evidence substantiating the selection criteria for children undergoing US after a UTI, particularly the age criterion of 2 years.

Following the emerging evidence, [7, 12, 13, 14, 15], our children's hospital was one of the first to discontinue using routine voiding cystourethrography after a UTI in children in the 2000s (Table S1). Subsequently, we allowed US to all children after a UTI to detect any possibly missed urinary tract abnormalities. This study investigated the prevalence of ultrasound findings in a cohort of all children undergoing US after a UTI and separately in different presumed risk groups, suggested by the current guidelines [2, 3, 4, 5, 16], at our hospital over a 12‐year period (2004–2015).

2. Methods

2.1. Study Design, Setting and Participants

In this register‐based cohort study, we included all children aged 0–16 years who underwent US of the urinary tract at Oulu University Hospital, Finland, between January 1, 2004 and December 31, 2015. All children treated at the hospital, referred from the outpatient clinic, or referred directly to the Department of Pediatric Radiology were included in this study. Children in whom the US was not performed because of a UTI episode, or if the US was performed more than 6 months after the UTI in children < 2 years and more than 1 year after UTI in children > 2 years were excluded. Additionally, children with a previously diagnosed pathology of the urinary tract, including abnormalities detected in prenatal US, were excluded from the study. The study protocol for this register‐based study was approved by Oulu University Hospital (74/2016). In accordance with Finnish legislation, individual informed consent was not required for this register‐based medical study. All reporting of the study was done according to STROBE guidelines. This study is a continuation of a previous study on radiological imaging after UTI from the same Finnish paediatric hospital including 1185 children treated between January 1, 1993 and December 31, 2003 [7].

2.2. Definition of Urinary Tract Infection

UTIs were diagnosed according to the Finnish national UTI treatment guideline in the healthcare units referring children to US [16]. In the guideline, paediatric UTI is defined as a significant growth of uropathogen in a urine sample in a symptomatic child. Accordingly, in the present study, significant bacterial growth was defined as any uropathogen growth in a suprapubic aspiration or uropathogen growth of > 10⁵ CFUs/mL in a urine sample gathered with another technique. The type of bacterial growth in the urine sample and the method of urine sample collection at the time of UTI diagnosis was gathered when available during the medical record review. In the study population, the urine sample was obtained via suprapubic aspiration in 402 (19.6%) children, via catheterization in 93 (4.5%) children, via clean catch urination in 805 (39.3%), via urine pad in 261 (12.7%), via urine bag in 109 (5.3%) children, and the method was undocumented in 380 (18.5%) children.

2.3. Data Sources

We surveyed the register of radiological examinations at Oulu University Hospital to identify eligible children based on the aforementioned selection criteria. Individual review of electronic medical records was then performed to acquire relevant data.

2.4. Ultrasound Imaging

A paediatric radiologist, a radiologist on call or a radiology resident supervised by a specialist performed the urinary tract US using the Philips iU22 (Philips Medical Systems, Amsterdam, The Netherlands) or Xario 200 (Canon Medical Systems Corporation, Ōtawara, Japan) ultrasound scanner and the recommended methods [17]. The radiologist chose the transducers most suitable for each patient; according to the patient's size, harmonic imaging and Doppler were used when needed. The bladder and kidneys were examined in axial and longitudinal planes, and kidney size, presence of parenchymal abnormalities, and enlargement of renal pelvis, calyces or ureters were reported. If possible, the post‐void evaluation was performed to assess the residual volume, bladder wall thickness, the renal pelvis and ureteric changes in the case of dilatation. During the study period, the cost of US in our hospital was €50–70 per examination.

2.5. Abnormalities of the Urinary Tract

The primary outcome of this study was the number and proportion of children with at least one abnormality in the urinary tract US after a UTI. Abnormalities were defined based on existing literature [18, 19, 20]. Briefly, US of the urinary tract was defined as abnormal if there was one or more of the following findings: upper urinary tract dilation, renal parenchymal abnormalities, an abnormally small kidney, urinary tract lithiasis or nephrocalcinosis, kidney agenesis or horseshoe kidney or bladder abnormalities. Detailed definitions of these abnormalities are provided in Table S2. The following ultrasound findings were considered normal: duplex kidney without dilatation or a large kidney (+2 Standard deviation [SD]). Ultrasound findings related to acute infection were ignored. Bladder wall thickening was considered a nonspecific ultrasound finding that can be caused by the acute infection and it was not included as an abnormality in this study.

2.6. Analysis of Abnormalities in Different Patient Groups

We reported the number and proportion of children with at least one abnormality separately for all children and the following group of children who are at a higher risk—infants (0–2 years), boys, children with pyelonephritis, children with non‐ Escherichia coli UTI, and children with recurrent UTI [2, 3, 4, 5, 16]. Pyelonephritis was defined as a UTI and either a fever > 37.9°C or C‐reactive protein levels of > 50 mg/L; the same criteria were used to define febrile UTI. Non‐ E. coli UTI was defined as the growth of any uropathogen other than E. coli in the urine sample. A full list of uropathogens included is provided in Table S3.

2.7. Bias

The following possible sources of bias were considered while designing the study. The possibility of selection bias was low in children with pyelonephritis because these children were both treated and actively screened in the study hospital, which was the only children's hospital in the area. However, it is possible that the likelihood of referral to US after a UTI diagnosis varied depending on the healthcare unit and practitioner, which increases the possibility of selection bias, in particular in older children with cystitis. To reduce information bias, all ultrasound examinations that had been originally reported as abnormal or borderline abnormal were reviewed by a paediatric radiologist during the study. The possibility of observer bias could not be fully excluded; however, we attempted to minimise its likelihood by using strict and standard definitions of urinary tract abnormalities. Lastly, to reduce confounding bias, we stratified the reporting of urinary tract abnormalities according to the possible risk groups (age, sex, diagnosis of pyelonephritis, uropathogen, recurrent UTI) [2, 3, 4, 5, 16]. Additionally, the logistic regression analysis was adjusted for sex and age.

2.8. Sample Size Calculation

We assumed 5%–10% incidence of an abnormal ultrasound finding in the children because during the study period, US was performed without strict selection criteria. Accordingly, we aimed for a sample size that may yield an estimated abnormality rate of 5% with a maximum deviation of 2%. The minimum sample size was calculated as 419.

2.9. Statistical Analysis

We reported the proportions of children with at least one urinary tract abnormality with 95% confidence intervals (CI) calculated using the Clopper–Pearson method for possible risk groups based on previous literature and guidelines [2, 3, 4, 5, 16]. Next, logistic regression was used to examine the association between abnormal ultrasound findings and age and sex, and logistic regression adjusted for age and sex was used to determine the association between abnormal ultrasound findings and the presence of pyelonephritis, or non‐ E. coli UTI, or recurrent UTI. Lastly, proportions with 95% CI were used to explore how well the current treatment guidelines detected the abnormalities in our study. There were no missing data regarding ultrasound results. Data were analysed using SPSS (IBM SPPS Statistics for Windows, version 29.0.1 Armonk, NY: IBM Corp.) and StatsDirect (version 3.3.6 http://www.statsdirect.com England: StatsDirect Ltd. 2024). Figures were drawn using Origin 2024b (OriginLab Corporation, Northampton, Massachusetts, USA) and SmartDraw (www.smartdraw.com SmartDraw LLC, The Woodlands, Texas, USA).

3. Results

3.1. Study Population

The initial study population comprised 4841 children who underwent US of the urinary tract during the study period (Figure 1). A total of 2524 children were excluded because UTI was not the indication for ultrasound imaging, including 457 children with prenatally detected urinary tract anomalies; 2317 children underwent US for UTI. Among these, 258 children were excluded because US was not performed within the set time limit after the UTI, eight children were excluded because of missing ultrasound data, and one child had a renal transplant. Eventually, the final study sample comprised 2050 children, most of whom were girls (75.7%). Among these, the US was considered normal for 1934 (94.3%) children (Table 1). Children with pyelonephritis were mainly younger than 2 years (n = 778, 68.2%). Children with cystitis and undefined levels of UTI were mainly older than 2 years (n = 315, 66.5% and n = 342, 78.6%, respectively).

FIGURE 1.

Flow chart of the study protocol.

TABLE 1.

Demographic and clinical characteristics of the study population (N = 2050).

	Children with normal ultrasound (n = 1934)	Children with abnormal ultrasound (n = 116)
Age, years mean (SD)	3.1 (3.2)	3.4 (4.1)
Sex
Boy, n (%)	455 (23.5)	43 (37.1)
Girl, n (%)	1479 (76.5)	73 (62.9)
Undefined	N/A	N/A
Level of urinary tract infection
Cystitis, n (%)	459 (23.7)	15 (12.9)
Pyelonephritis, n (%)	1070 (55.3)	71 (61.2)
Undefined, n (%)	405 (20.9)	30 (25.9)
Uropathogen
Escherichia coli , n (%)	1401 (72.4)	61 (52.6)
Non‐ E. coli , n (%)	171 (8.8)	24 (20.7)
Unknown, n (%)	362 (18.7)	31 (26.7)
Follow‐up care due to abnormality in ultrasound screening a
≥ 1 visits to paediatrician, n (%)	N/A	28 (24.1)
≥ 1 visits to paediatric urologist, n (%)	N/A	73 (62.9)
≥ 1 visits to paediatric nephrologist, n (%)	N/A	7 (6.0)
Cystoscopy or urological surgery, n (%)	N/A	32 (27.6)
Only cystoscopy, n (%)	N/A	5 (4.3)
Surgical operation, n (%)	N/A	27 (23.3)

Note: All urinary tract ultrasound scans were performed by radiologists in specialised healthcare at paediatric hospital.

^a Same child can have visits to multiple different specialists as part of follow‐up care.

3.2. Abnormalities of the Urinary Tract in the Entire Cohort

At least one abnormality of the urinary tract was found in 116 (5.7%; 95% CI: 4.7%–6.7%) children, the most common being upper urinary tract dilation (n = 76, 3.7%) children (Table 2). Thirteen children (0.6%) had renal parenchymal abnormalities, and 32 (1.6%) had bladder abnormalities. Other urinary tract abnormalities were rare.

TABLE 2.

The number and proportion of children with abnormalities found in the ultrasound screening after a urinary tract infection stratified for different risk groups.

Ultrasound screening results	All children (N = 2050)	Age < 2 years (n = 1030)	Boys (n = 498)	Girls (n = 1552)	Children with pyelonephritis (n = 1141)	Non‐ E. coli UTI (n = 195)	Recurrent UTI (n = 96)
Normal, n (%)	1934 (94.3)	965 (93.7)	455 (91.4)	1479 (95.3)	1070 (93.8)	171 (87.7)	84 (87.5)
Abnormal, a n (%) [95% CI of proportion]	116 (5.7) [4.7–6.7]	65 (6.3) [4.9–8.0]	43 (8.6) [6.3–11.5]	73 (4.7) [3.7–5.9]	71 (6.2) [4.9–7.8]	24 (12.3) [8.0–17.8]	12 (12.5) [6.6–20.8]
Upper urinary tract dilatation, n (%)	76 (3.7)	54 (5.2)	33 (6.6)	43 (2.8)	54 (4.7)	21 (10.8)	3 (3.1)
Renal parenchymal abnormalities, n (%)	13 (0.6)	6 (0.6)	5 (1.0)	8 (0.5)	8 (0.7)	2 (1.0)	0 (0.0)
Abnormally small kidney, n (%)	3 (0.1)	1 (0.1)	0 (0.0)	3 (0.2)	3 (0.3)	0 (0.0)	0 (0.0)
Urinary tract lithiasis or nephrocalcinosis, n (%)	4 (0.2)	1 (0.1)	1 (0.2)	3 (0.2)	0 (0.0)	1 (0.5)	1 (1.0)
Kidney agenesis or horseshoe kidney, n (%)	2 (0.1)	2 (0.2)	1 (0.2)	1 (0.1)	1 (0.1)	0 (0.0)	1 (1.0)
Bladder abnormalities, b n (%)	32 (1.6)	8 (0.8)	7 (1.4)	25 (1.6)	13 (1.1)	3 (1.5)	8 (8.3)
Urethrocele	10 (0.5)	7 (0.7)	2 (0.4)	8 (0.5)	5 (0.4)	2 (1.0)	0 (0.0)
Diverticulum	1 (< 0.1)	1 (0.1)	1 (0.2)	0 (0.0)	1 (0.1)	1 (0.5)	0 (0.0)
Dilated posterior urethra	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Abnormal residual	21 (1.0)	0 (0.0)	4 (0.8)	17 (1.1)	7 (0.6)	0 (0.0)	8 (8.3)
Odds ratio of abnormal ultrasound c (95% CI) [p‐value]	N/A	0.99 (0.65–1.52) [0.97]	2.20 (1.45–3.34) [< 0.001]	0.45 (0.30–0.69) [0.04]	1.31 (0.87–1.96) [0.19]	2.96 (1.77–4.96) [< 0.001]	2.65 (1.36–5.19) [0.004]

Abbreviations: CI, confidence interval; UTI, urinary tract infection.

^a One or more of the listed abnormalities.

^b There were 72 children with bladder wall thickening in the cohort. However, this was not considered a significant abnormality per our study protocol.

^c Children less than 2 years old compared to children at least 2 years old, boys compared to girls, girls compared to boys, children with pyelonephritis compared to children with cystitis or undefined level of UTI, non‐ E. coli UTI compared to E. coli UTI, children with recurrent UTI compared to those without recurrent UTI. Analyses for pyelonephritis, non‐ E. coli UTI and recurrent UTI were adjusted for the age and sex of the children.

3.3. Abnormalities of the Urinary Tract in Children With Pyelonephritis

Pyelonephritis was the indication for US in 1141 children (Table 2), which revealed an abnormal finding in 71 (6.2%, 95% CI: 4.9%–7.8%) children. Upper urinary tract dilatation was the most common abnormality (n = 54, 4.7%); renal parenchymal abnormalities were found in 8 (0.7%) and bladder abnormalities in 13 (1.1%) children with pyelonephritis. Other categories of abnormalities were rare. In children with pyelonephritis, 70.4% of the abnormal ultrasounds were found in children younger than 2 years (Figure 2). Pyelonephritis was not statistically significantly associated with abnormal US (OR: 1.31, 95% CI: 0.87–1.96).

FIGURE 2.

Abnormal and normal urinary tract ultrasound imaging results after a urinary tract infection in children during childhood.

3.4. Abnormalities of the Urinary Tract in Infants

Out of the 116 cases of children with urinary tract abnormalities, 65 (56%) were observed in children younger than 2 years of age, making up for 6.3% (95% CI: 4.9%–8.0%) of all children under 2 years of age in our study. After infancy, the cumulative proportion of children with abnormalities increased throughout childhood (Figure 2). The age < 2 years was not statistically significantly associated with an increased risk of abnormality in the US compared to age > 2 years (odds ratio, OR: 0.99, 95% CI: 0.65–1.52) (Table 2).

3.5. Abnormalities of the Urinary Tract in Boys

Altogether, 498 boys underwent US of the urinary tract during the study period (Table 2). Ultrasound was considered abnormal for 43 (8.6%, 95% CI: 6.3%–11.5%) boys. Upper urinary tract dilatation was the most common abnormality (n = 33, 6.6%), followed by bladder abnormalities (n = 7, 1.4%) and renal parenchymal abnormalities (n = 5, 1.0%); other categories of abnormalities were rare. Most boys (85.3%) were younger than 2 years, and most of the abnormal ultrasounds (81.4%) were found in boys younger than 2 years (Figure 2). Sex was statistically significantly associated with an abnormal US in this study (boys versus girls: OR: 2.20, 95% CI: 1.45–3.34).

3.6. Abnormalities of the Urinary Tract in Children With Non‐ E. coli UTI

A total of 195 children had non‐ E. coli UTI (Table 2). US was considered abnormal in 24 (12.3%, 95% CI: 8.0%–17.8%) children. Upper urinary tract dilatation was the most common abnormality found in 21 (10.8%) children; only six children had other abnormalities. In this subgroup, 66.7% of abnormal ultrasounds were observed in children younger than 2 years (Figure 2). Compared to UTI caused by E. coli , non‐ E. coli UTI was associated with a statistically significantly increased risk of abnormal ultrasound findings (OR: 2.96, 95% CI: 1.77–4.96).

3.7. Abnormalities of the Urinary Tract in Children With Recurrent UTI

Ninety‐six children in this study had recurrent UTI (Table 2). US was considered abnormal for 12 (12.5%, 95% CI: 6.6%–20.8%) children. Bladder abnormality was the most common abnormality (n = 8, 8.3%), whereas 3 (3.1%) children had upper urinary tract dilatation. Only one abnormality was found before the age of 2 years (Figure 2). Recurrent UTI was associated with a statistically significantly increased risk of abnormality in the ultrasound imaging (OR: 2.65, 95% CI: 1.36–5.19).

3.8. Follow‐Up and Surgery Needed in Children With an Abnormal Ultrasound

Overall, 91 children (78.4% of the children with abnormality, 4.4% of all children) needed clinical follow‐up because of abnormal ultrasound findings (Figure 3). Thirty‐two children (27.6% of children with abnormal ultrasound, 1.6% of all children) underwent cystoscopy (n = 5) or urological surgery for upper urinary tract obstruction or urethral valve (n = 27). Three of the children who only underwent cystoscopy were younger than 2 years, and two were older than 2 years. In the whole cohort, the estimated cost for finding one abnormality requiring cystoscopy or urological surgery was €4060 (approximately $4300). In children older than 2 years, 33 (64.7% of the children with abnormal ultrasound, 3.2% of all children) needed clinical follow‐up, and 11 (21.6% of the children with abnormal ultrasound, 1.1% of all children) needed cystoscopy or urological surgery. A separate electronic medical records check revealed that 11 (52.4%) of children with abnormal residual had symptoms (enuresis, pollakisuria or constipation) linked to bladder and bowel dysfunction (BBD) and 9 (81.8%) of these children had received medication aimed at treating BBD.

FIGURE 3.

Outcomes of children with abnormalities in ultrasound imaging after a urinary tract infection.

3.9. Performance of the Current Treatment Guidelines in the Study Cohort

The AAP guideline targeted only at children 2–24 months old would have missed 54% (95% CI: 41%–66%), the NICE guidelines 45% (95% CI: 36%–54%), and the EAU/ESPU guidelines 30% (95% CI: 19%–42%) of the children with abnormal US in this cohort (Table 3). On the other hand, the Asian guideline, with no specific age criteria, would not have missed any abnormalities.

TABLE 3.

Proportion of children with abnormal urinary tract ultrasound that would not have been found with the criteria set by different guidelines.

Guideline criteria	AAP	NICE	EAU/ESPU	ASIAN
Main criteria for selecting children	Age 2–24 months (Febrile UTI) a	Age < 6 months	Infants (Febrile UTI)	No specific age criteria (Febrile UTI)
Imaging also recommended:
Recurrent UTI	ND	YES	ND	ND
Atypical UTI b	ND	YES	ND	ND
Proportion of children with abnormal ultrasound in this study the guideline criteria would have missed c (95% CI)	54% d (41%–66%)	45% e (36%–54%)	30% f (19%–42%)	0%

Abbreviations: AAP, The American Academy of Pediatrics; CI, confidence interval; EAU/ESPU, European Society for Paediatric Urology; ND, not defined; NICE, The National Institute for Health and Care Excellence; UTI, urinary tract infection.

^a Febrile UTI was defined as UTI with either fever > 37.9°C or c‐reactive protein level of > 50 mg/L.

^b Non‐ E. coli UTI, seriously ill child, or child not responding to treatment.

^c Due to some children missing fever and C‐reactive protein data, only children with confirmed febrile UTIs were included in the analysis if febrile UTI was used as criteria in the guideline. There were 116 children with abnormal ultrasound, and 71 of these children had confirmed febrile UTI.

^d Proportion of children with abnormal ultrasound, who were either < 2 months old or > 2 years old and diagnosed with pyelonephritis.

^e Proportion of children with abnormal ultrasound, who were > 6 months old and had no recurrent UTI or non‐ E. coli UTI.

^f Proportion of children with abnormal ultrasound, who were > 2 years old and diagnosed with pyelonephritis.

4. Discussion

This study included a large cohort of 2050 children aged 0–16 years, who were subjected to US after a UTI without strict selection criteria. This US detected urinary tract abnormalities in 6% of children. Boys, children with a non‐ E. coli UTI, and children with a recurrent UTI appeared to have a higher risk for urinary tract abnormalities. Currently, most guidelines recommend US for infants (< 2 years of age) after a UTI. In this study, however, we found abnormalities throughout childhood, meaning that a significant proportion (44%) of the children with at least one abnormality would have been missed if US were not performed for children above 2 years of age. In children and adolescents with pyelonephritis, 30% of children with at least one abnormality would have been missed if US were performed only for children younger than 2 years.

The results of the present study are in line with a recent meta‐analysis on US after the first febrile UTI, that included 29 heterogeneous studies with sample sizes ranging from 66 to 820 children [21]. The meta‐analysis showed that clinically significant abnormalities were found in 3% of children of all ages [21]. Furthermore, the prevalence of abnormalities was comparable in studies of children younger than 2 years and those including older children [21]. The present study included 2050 children, making it one of the largest to date. Based on these findings, we believe that the age criterion of the present treatment guidelines may be too restrictive and needs reconsideration.

US performed according to the AAP recommended protocols is estimated to cost $ 803 000 per quality adjusted life year gained in the United States [22]. Thus, the clinical benefits of US should be carefully evaluated. We acknowledge that the availability and cost of US vary between countries and hospitals, which must be considered when choosing the criteria for imaging in the guidelines. In the context of the present study, US was easily available and relatively cheap, resulting in an estimated cost of $4300 per one found urinary tract abnormality requiring cystoscopy or urological surgery.

We found that boys had a significantly increased risk of any urinary tract abnormality. Most treatment guidelines do not acknowledge boys as a separate risk group in the selection criteria for ultrasound imaging [2, 3, 4, 5]. Previous literature shows that renal parenchymal defects in radiological imaging after a UTI are more commonly found in boys [23, 24]. The present study was conducted in Finland, where most boys are uncircumcised, which is a known risk factor for UTI in boys [25]. This restricts the generalisability of our results to populations with a high proportion of circumcised boys. However, our findings suggest that boys, at least in mainly uncircumcised populations, are at increased risk of urinary tract abnormality, and this should be considered as a criterion for ultrasound imaging in the treatment guidelines.

Most guidelines recommend US for children after febrile UTIs [2, 4, 5]. In this study, pyelonephritis (defined as fever and/or elevated CRP value) was not found to be a statistically significant predictor of urinary tract abnormalities. We assume that not all children with cystitis were sent to US from outpatient care in this cohort study. Some children had missing clinical data on fever and C‐reactive protein levels. Thus, we could not reliably classify the level of UTI for every child included in the study. Because of possible selection and classification bias, we do not suggest removing febrile UTI as a selection criterion for US.

Non‐ E. coli UTI has been previously suggested to be a possible risk factor for detecting urinary tract abnormalities in US [26, 27]. In our study, non‐ E. coli UTI was associated with an increased risk of urinary tract abnormalities; however, the number of children with non‐ E. coli UTI in the study was low. Some studies have suggested that US should be performed only after recurrent UTI, as this approach would not miss significant abnormalities and would reduce the number of imaging studies needed [22, 27, 28]. In this study, recurrent UTI seemed to predict urinary tract abnormality as well, but the total number of children with recurrent infections was low.

In this study, bladder abnormalities were more commonly found in older children than infants. Of the children with abnormal residual, more than half had BBD symptoms and medication aimed at treating BBD was commonly administered. The prevalence of BBD has been reported to be as high as 54% in children who experience UTIs, and it has been suggested that toilet‐trained children suffering from UTIs could benefit from active screening for BBD [29]. In a recent meta‐analysis, however, there was no statistically significant association found between daytime or night enuresis and the occurrence of UTIs [30]. In the same meta‐analysis, obstipation did not reach statistical significance as a risk factor for recurrent UTI, but there appeared to be an association between infrequent voiding and the risk of recurrent UTI [30]. Further studies are needed to fully perceive how significant finding abnormal post‐void residual is in children after a UTI.

It is noteworthy that we excluded children in whom urinary tract abnormality was already detected in the prenatal ultrasound. A recent meta‐analysis reported that normal prenatal ultrasound gives a negative predictive value of 85% for significant abnormalities of the urinary tract after a UTI [31]. In Finland, prenatal ultrasound is offered to all pregnant women as part of free maternity health clinic services; thus, most children have been screened by a prenatal ultrasound before birth. The frequent use of high‐quality prenatal US in our study population likely decreased the proportion of abnormal ultrasound findings in infants after an UTI episode.

The main strength of our study was the use of US for children of all ages, enabling us to study urinary tract abnormalities throughout childhood in a population‐based cohort. Furthermore, we used a large sample compared to previous studies on this subject [21]. It should also be considered a strength that all ultrasound examinations were performed in the same university hospital by radiologists, allowing for uniform methods and quality of ultrasound examinations throughout the study. Finally, we also reported the number of surgical operations needed after an abnormal ultrasound.

However, there are some limitations. Regarding the generalisability of our results, the study was conducted in a high‐income country with high‐quality prenatal care, where most of the boys were uncircumcised. The main limitation of the study design was selection bias regarding the study population. It is possible that especially some older children with cystitis were not referred for US from outpatient care, even though it was allowed. Thus, the prevalence of urinary tract abnormalities, especially in children with non‐severe UTIs, may have been overestimated. As US was actively performed in all children after pyelonephritis, the selection bias is less likely in this subgroup. Thus, we have given the results separately in children with pyelonephritis to help the interpretation of the results. Lastly, the clinical data on the level of the UTI and microbiological results were missing for some children.

5. Conclusion

This cohort study revealed that urinary tract abnormalities are found in US after a UTI throughout childhood. Restricting US after a pyelonephritis to children younger than 2 years may result in missing 30% of children with at least one abnormality. Therefore, we suggest considering US after the first UTI episode for all children with pyelonephritis, including those older than 2 years. A prospective longitudinal study comprising all children with a diagnosed UTI is needed in the future.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1

Data Availability Statement

Due to data protection regulations, only restricted data will be shared upon a reasonable request for clinical study purposes. Requests with a full research protocol can be sent via email to the principal investigator Terhi Ruuska‐Loewald (terhi.ruuska-loewald@oulu.fi).