Article Text
PDF
Abstract
Background Imaging is recommended for selected children following urinary tract infections (UTIs) to look for actionable structural abnormalities. Non-E. coli is considered high risk in many national guidelines, but evidence is mainly drawn from small cohorts from tertiary centres.
Objective To ascertain imaging yield from infants and children <12 years diagnosed with their first confirmed UTI (pure single growth >100 000 cfu per ml) in primary care or an emergency department without admission stratified by bacteria type.
Design, setting, patients Data were collected from an administrative database of a UK citywide direct access UTI service between 2000 and 2021. Imaging policy mandated renal tract ultrasound and Technetium-99m dimercaptosuccinic acid scans in all children, plus micturating cystourethrogram in infants <12 months.
Results 7730 children (79% girls, 16% aged <1 year, 55% 1–4 years) underwent imaging after first UTI diagnosed by primary care (81%) or emergency department without admission (13%). E. coli UTI yielded abnormal kidney imaging in 8.9% (566/6384). Enterococcus and KPP (Klebsiella, Proteus, Pseudomonas) yielded 5.6% (42/749) and 5.0% (24/483) with relative risks 0.63 (95% CI 0.47 to 0.86) and 0.56 (0.38 to 0.83)), respectively. No difference was found when stratified by age banding or imaging modality.
Conclusion In this largest published group of infants and children diagnosed in primary and emergency care not requiring admission, non-E. coli UTI was not associated with a higher yield from renal tract imaging.
- Emergency Care
- Microbiology
- Primary Health Care
- Nephrology
- Infectious Disease Medicine
Data availability statement
Data are available upon reasonable request. On reasonable request aggregated non patient identifiable data can be shared.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC?
Non-E. coli urinary tract infections (UTIs) have historically been associated with high risk of underlying renal tract abnormalities.
This association comes from several small studies of inpatient mainly in tertiary hospitals.
Many national guidelines recommend imaging after a first non-E. coli UTI in children, regardless of severity.
WHAT THIS STUDY ADDS?
After a first UTI treated in primary care or in emergency department, the prevalence of structural renal tract abnormalities were no higher in those with non-E. coli vs E. coli infections.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY?
These observations require validation by further studies, ideally without pad urine samples.
Children with a first non-E. coli UTI who are not hospitalised may be evaluated in the same manner as children with E. coli UTIs.
Introduction
Approximately 2%–6% of acutely ill children <5 years old presenting to primary care in the UK are diagnosed with a urinary tract infection (UTI).1 2 Imaging is recommended for selected children following UTI to look for structural abnormalities, such as vesicoureteric reflux (VUR), which may be actioned to reduce recurrence and consequential risk of long-term kidney damage.
Non-E. coli UTI is considered high risk for underlying structural abnormalities in many national guidelines (including but not limited to: UK National Institute for Health and Care Excellence (NICE),3 multidisciplinary Swiss consensus,4 Kidney Health Australia—Caring for Australasians with Renal Impairment (KHA-CARI)5 and Italian Society Of Paediatric Nephrology6), leading to more invasive imaging. However, published evidence for increased structural abnormalities detected on imaging is limited to small cohorts of children diagnosed and treated as in-patients rather than in primary or emergency department without admission.
The North-East of England has historically been a strong proponent of universal screening after the first UTI and has continued until recently.7 We present a unique, large contemporary data set of imaging results after first E. coli vs non-E. coli UTI in children <12 years old diagnosed and treated in primary or emergency care without hospital admission.
Method
This was a retrospective cohort study involving all children <12 years old referred with a clinical diagnosis of UTI to the Newcastle UTI direct access service between 2000 and 2021 (21 years). The aim of the study was to ascertain imaging yield after a first confirmed UTI in primary care, or an emergency department without admission, stratified by bacteria type.
Data related to referrals and imaging outcomes of these children were collected contemporaneously in an administrative database (FileMaker Pro, Claris International) to track and manage referrals. The UTI direct access service accepts referrals of children requesting investigation after UTI from primary care and the emergency department for the city of Newcastle Upon Tyne, UK (population 45 000 aged 0–14 years in 20118), and surrounding areas if requested.
During the study period, imaging policy for a first confirmed UTI was ultrasound and dimercaptosuccinic acid (DMSA) scans in all children, plus micturating cystourethrogram (MCUG) in infants <12 months. DMSA scans were deferred for a minimum of 3, later 4, then 6 months from resolution of UTI. From 2015, only children <8 years were imaged after a first UTI and <5 years from 2020. Children were accepted for imaging if symptoms were accompanied by a pure single growth ≥105 CFU/mL from any urine collection method. Urine culture result was verified by our service administrator by cross-referencing with the region wide electronic laboratory system (Sunquest Integrated Clinical Environment (ICE)). Children who underwent previous kidney imaging for any reason were excluded from analysis. The policy of our emergency department is diagnosis and treatment within 4 hours of attendance—patients requiring longer management are admitted to hospital paediatric care and were excluded from this cohort. We were unable to obtain sufficiently robust or comprehensive data for hospitalised children for a comparison group.
Although the service and database was set up in 1998, we excluded patients referred before 1 January 2000 to allow a period of settling in to reduce patient selection bias. Also excluded were referrals after 31 December 2021 as not all had completed their investigations.
Data available included sex, age, reported symptom of fever, urine collection method, urine culture result and imaging result. Bacterial species were grouped by the database into ‘E. coli’, ‘Enterococcus’, ‘KPP’ (Klebsiella, Proteus, Pseudomonas) or ‘others’. KPP infections were aggregated in the database, as the limited literature published at the inception of the service suggested that they were most frequently associated with imaging abnormalities.9 10 Records of infections with bacteria classified as ‘others’ in the database were further interrogated in ICE, with data available from 2008 to present.
MCUG results were classified into ‘no’, ‘low grade’1 2 or ‘high grade’3–5 VUR as per International System of Radiographic Grading of VUR, as well as any significant bladder findings requiring intervention.11 DMSA reports were classified as ‘scarred’ or not scarred regardless of severity of scarring. Ultrasound scans (USS) were classified as ‘normal’ or ‘abnormal’ regardless of severity of abnormality. USS with uncertain findings were classified as abnormal. Variants on DMSA scan without scarring and seen on USS such as unilateral, horseshoe or ectopic kidneys were classified as abnormal USS appearance and non-scarred DMSA. In the clinical service, all abnormal or uncertain scans were reviewed at a weekly regional multi-professional nephro-urology radiology meeting.
Results are expressed as relative risk (RR) with 95% CI. χ2 compared expected and observed values with significance set at p<0.05. Data was analysed using Excel V.2206 (Microsoft Corporation). This project was registered with our institutional clinical effective and audit department as service evaluation (ref 13950). No patient identifiable data were collected and ethical approval was not considered necessary as per NHS Health Research Authority definition.
Results
As of May 2022, our UTI direct access administrative database contained information on 18 114 referrals from 1998 of whom 7846 met the inclusion criteria (figure 1). A further 116 were excluded because microbiology reports did not differentiate clearly into the defined bacterial subgroups to allow for analysis (67 coliforms (K.E.S.), 11 coliforms and 38 unknown other). This left 7730 cases for analysis.
Patient extraction data for study. UTI, urinary tract infection.
Baseline characteristics are outlined in table 1. Children were grouped into: <1 year (n=1219), 1–4 years (n=4281) and 5–11 years (n=2230). Apart from <1 year olds, girls predominated (n=6121, 79%). Children were mainly diagnosed and treated at their own GP practice (n=4754, 62%) or at emergency department (n=786, 10%) although data was missing in 1867 (24%). In infancy, urine was collected mainly by urine pads (894/1219, 73%) and by clean catch for older age groups.
Baseline characteristics of infants and children who underwent kidney imaging following first confirmed UTI diagnosed and treated in primary or emergency care
Imaging revealed an abnormality in 639/7730 (8.3%) of all children. This ranged from 274/1219 (23%) in <1 year to 244/4281 (5.7%) in 1–4 years and 121/2230 (5.4%) in 5–11 year olds. As <1 year olds had a different imaging strategy, when only USS and DMSA were considered, abnormalities were found in 82/1210 (6.8%) infants. One posterior urethral value was found on MCUG after an E. coli UTI.
Overall, 1345 (17.4%) were non-E. coli UTI with Enterococcus predominating (n=749, 9.6%). Bacteria comprising the ‘other’ group are listed in table 2. There were no significant differences between urine collection method and bacterial group isolated in any of the age groups (online supplemental table S1). The RRs of finding scan abnormalities were not significantly higher in any of the non-E.coli UTI groups as a whole nor when stratified by age or imaging modality (table 3). Indeed, Enterococcus and KPP groups had lower risks of scan abnormalities compared with E. coli, RR 0.63 (95% CI 0.47 to 0.86) and RR 0.56 (0.38 to 0.83), respectively (figure 2). In subgroup analysis, this remained significant in the 1-year to 4-year age group.
Supplemental material
Relative risk of an abnormality on kidney imaging following a first UTI in infants and children <12 years old stratified by bacterial type. Error lines indicate 95% CI. UTI, urinary tract infection.
Bacteria isolated in the ‘other’ group (n=114)
Subgroup analysis divided by imaging modality and age band
Where data was available, 2201/5629 (37.1%) reported symptom of fever. Fever was less prevalent in non-E. coli (359/1020 (35%)) vs E. coli (1842/4609 (40%), p=0.005) UTI. Subgroup analysis showed this remained true for <1 year olds and 5–11 year olds (online supplemental table S2). When only children with fever symptom were analysed, non-E. coli UTI continued to show lower risks of scan abnormalities compared with E. coli, RR 0.59 (95% CI 0.40 to 0.88).
Discussion
Using a large administrative database, we did not demonstrate an increased risk of structural abnormalities with non-E. coli vs E. coli UTI for infants and children who were diagnosed and treated in primary care or in ED without admission.
The first report suggesting an association with structural abnormalities and non-E. coli UTI was a case–control Finnish hospital in-patient study of children admitted 1980–1994.10 Using catheter and suprapubic aspirate samples, just 26 children each with Proteus, Klebsiella or Enterococcus UTI and 16 with coagulase-negative staphylococcal UTI were compared with 92 E. coli UTI controls. In every group except proteus, significantly increased imaging abnormalities were found, and many proceeded to surgical interventions. Our result contrasts with many smaller studies published since 1999 showing increased structural abnormalities (table 4) after non-E.coli compared with E. coli UTI, but all of these were conducted in hospital settings, mostly tertiary hospitals. This list is not exhaustive. Comparing studies was difficult as inclusion criteria were heterogeneous, recurrent UTI rates varied, imaging protocols differed, as did result reporting layout and even comparison groups. For example, one tertiary hospital in-patient study in <18 year olds in Israel found greater abnormalities in Enterococcus UTI vs all other gram-negative UTI.12
Review of published studies since 1999 in children comparing imaging outcome after E. coli vs non-E. coli UTI
There are study limitations. We used an administrative database managed by multiple non-clinical operators which led to input errors and missing data, but our methodology was no different to the other published studies, comprising retrospective clinical cohorts. Misclassification bias, unmeasured confounding and changing eligibility over time are likely, although referrals to the UTI service have remained constant over the study period. To our knowledge, our cohort is larger than the combined total of previous studies and uniquely excludes children who were admitted to hospital.
It is conceivable that non-E. coli UTI may be acutely sicker and more admitted to hospital. We found symptoms of fever was slightly less frequent in non-E. coli UTI. A comprehensive comparison group over the same geography and timeframe of hospitalised children would have been useful but our current databases does not have that information.
Outside hospital settings, clean catch urine sampling should be the gold standard collection method. Although our cohort had a high prevalence of pad specimens, this reflects real-life challenges in UK primary care. In the multi-centre DUTY study involving 225 GP surgeries and 7163 children <5 years old in South England and Wales, 50% of urine samples collected were pad samples and 41% were collected in the child’s home.2 Pads appear to be a convenient urine sampling method, preferred by parents and staff, but outside clinical trial settings they have high contamination rates.13 14 Replacing the pad for clean catch is challenging in busy hospital settings and unrealistic in primary care despite ultimately being more cost and time effective for the health economy.13 15 Reassuringly, we did not find non-E. coli bacteria over-represented in pad collection samples compared with clean catch. Managing children with recurrent symptoms attributed to the urinary tract in primary care is difficult and this is reflected in the large numbers of referrals into our service of low and mixed growth from pad samples; these patients were excluded from the study.
Children presenting with a first UTI due to non-E. coli organisms have been reported to have higher recurrence risk, but this finding should similarly be interpreted with caution, as that study was from children treated in hospital settings and again may represent a higher risk group.16 In one study, persistence of acute DMSA pyelonephritis changes appears to be more common in non-E. coli UTI, but it would be rare for acute DMSA scans to be done on acutely unwell children who did not require hospitalisation.17 In post hoc analysis of the prospective longitudinal studies, Randomized Intervention for Children With Vesicoureteral Reflux trial (RIVUR) and Careful Urinary Tract Infection Evaluation (CUTIE) studies in which 482 young children with febrile UTI had DMSA scans performed at study start and 24 months later, non-E. coli UTI patients had similar duration of fever and acquired scarring when compared with E. coli UTI.18 With increasing understanding of the human genome and metagenome, biomarkers may become a more useful tool in stratifying which children have host vulnerabilities for future pyelonephritis.19
National UTI guidelines already stratify children who are significantly ill or have recurrent UTIs for more imaging.3–6 This study requires validation by an independent non-hospitalised cohort, ideally without pad urine samples. If confirmed, we recommend that national guidelines are modified so first non-E. coli UTI in children who did not require hospital admission nor have additional risk factors are not subjected to excessive imaging.
Data availability statement
Data are available upon reasonable request. On reasonable request aggregated non patient identifiable data can be shared.
Ethics statements
Patient consent for publication
Acknowledgments
We wish to thank our UTI service administrator and specialist nurse Angela Mufford and Tara Craig for their dedication in improving the service.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Twitter @YincentTse, @owens_stephen
Contributors YT conceived the article, produced the initial draft and is acting as the guarantor; CP, manual search ICE data. All authors with their different skills and unique perspective inputted into initial draft, revised and approved the final article. Since 2022, YT has been the service lead for the Newcastle Direct Access UTI service.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.