Diagnosing Pediatric Urinary Tract Infections

Although diagnosing urinary tract infection (UTI) in a pediatric population can be challenging, pre-test probability can be guided by specific risk factors in the history and physical examination. Among low-risk children, further laboratory testing can be deferred in favor of a conservative approach.
The Setup

You’ve spent your morning locked in administrative meetings. After being serenaded at length by the budgetary mantra to cut costs while maintaining high-quality patient care, the chaos of your subsequent afternoon ED shift is almost a welcome contrast. The next chart in the rack, though, brings back memories of the morning’s subject matter…


Concerned parents have brought their 4-year old son to the ED with just over 24-hours of fever. History does not reveal a specific infectious focus. There is no vomiting, diarrhea, abdominal discomfort, or back pain. Hydration has been maintained. When asked about dysuria the patient giggles and hugs his mother instead of answering. His parents do not give any history of their child having complained of pain with urination. He is previously healthy, and uncircumcised. The nurse asks if you would like a urine sample sent to the lab. The dollar figures associated with unnecessary testing involuntary come to mind.

The Background
Urinary tract infections in children account for an estimated 5% to 15% of pediatric emergency department visits in the United States. The diagnosis can be difficult to establish as the signs and symptoms at the time of presentation can be highly variable and nonspecific, especially in younger, non-verbal patients. While guidelines exist (NICE, Cincinnati, CKS) that provide physician guidance to the diagnostic process, there is often emphasis on urinary cultures as the gold standard diagnostic test. The difficulty, however, is that these culture results are usually unavailable at the time a child presents to an emergency department. Delays in the diagnosis of pediatric UTI can result in renal scarring which can ultimately lead to hypertension and end-stage renal disease. A pragmatic approach to the diagnostic process in the emergency department that emphasizes appropriate early identification, balanced against the costs of indiscriminate testing, is required.

The Question

Does this child have a urinary tract infection? Is there a way to rationally approach the diagnosis while minimizing unnecessary testing?

The Evidence
Shaikh N, Morone NE, Lopez J, Chianese J, Sangvai S, D’Amico F, Hoberman A, Wald ER. “Does this child have a urinary tract infection?” JAMA. 2007 Dec 26;298(24):2895-904.

The Results

This JAMA Rational Clinical Examination systematically evaluated pediatric UTI diagnostic test characteristics for commonly obtained elements of the medical history, physical examination, and laboratory analysis. Based on prospective data provided by the 12 identified articles, the authors were able to calculate likelihood ratios for each step in the diagnostic process. These are mapped as suggested algorithms to guide diagnosis while minimizing testing in situations where UTI is clinically unlikely (pre-test probability < 2%). For infants (ages3-24 months), the following risk factors are considered:
  • History of UTI
  • Temperature > 39 C
  • Ill appearance
  • Suprapubic tenderness
  • Fever >24h
  • Nonblack race

In uncircumcised male infants, the baseline risk of UTI is around 6%. The presence of any of the aforementioned UTI risk factors is sufficient to raise this probability to between 10% and 25%, mandating further urinalysis and culture. In the absence of all the above factors, the risk falls to under 2%, and can be managed conservatively with reassessment in 24-hours.

In circumcised males the baseline risk is lower (around 1%) so two or more risk factors (or suprapubic tenderness alone) need to be present before further testing would be required. For female infants, the presence of any UTI risk factor is sufficient to raise the probability of UTI above the 2% threshold supporting urinalysis.

Click on image to view high-res PDF


Reprinted from Annals of Emergency Medicine, “Does This Child Have a Urinary Tract Infection?”, May 2009, Rupinder Singh Sahsi and Christopher R. Carpenter, copyright 2009, with permission from Elsevier

The algorithm is different for verbal children who present with urinary or abdominal symptoms. In circumcised males with a low baseline probability of UTI (<1%) urinalysis and culture is only recommended in the presence of multiple signs and symptoms of UTI. For females and uncircumcised males, the presence of dysuria/frequency OR the presence of abdominal pain, back pain, or new-onset incontinence is needed before diagnostic testing is mandated. In all other cases, UTI is considered highly unlikely. In this situation, alternative diagnoses should be contemplated while ensuring appropriate follow up. The Caveat

While constructed in a logical and appropriate manner, this article’s Baysean approach is highly dependent upon the available evidence uncovered by their systematic search algorithm. Future prospective evaluation is needed before these recommendations can be deemed sufficiently validated to guide clinical decisions in the form of a stand-alone decision aid.

The Outcome

In this case, our young patient’s baseline risk of UTI as an uncircumcised male is estimated at around 8%, and would be increased by the presence of specific symptoms. In the absence of urinary frequency or dysuria, and lacking the other “higher-risk” symptoms (abdominal pain, back pain, or new-onset incontinence) the algorithm would suggest that UTI is unlikely. Since a urinalysis is unlikely to be diagnostic, you opt for symptomatic management of a likely viral illness. After ensuring appropriate follow-up is available you pat yourself on the back for achieving a small (but important and evidence-based!) fiscal victory.

Rupinder Sahsi, BSc MD is a faculty member and Technology Director of BEEM (Best Evidence in Emergency Medicine) and is an assistant clinical professor of emergency medicine at McMaster University. He practices in the emergency departments of Kitchener-Waterloo, Ontario, Canada.


  1. Arthur Deininger, M.D. on

    You might have mentioned the female infant with an agglutinated labia, which increases the risk for UTI. Always spread open the labia to check for this. Otherwise, you won’t find it.

  2. A good point, Dr. Deininger. That particular risk factor wasn’t amongst the common ones identified by the systematic literature review. I suspect its contribution would probably be captured under the “history of UTI” risk factor.

    I wonder if there’s reasonable objective evidence to support this factor that many of us already suspect.

  3. Saving an unnecessary urine culture will cut some costs, but will you do a simple and cheap urine dip analysis? Even if it’s just to show the parents that the urine does look clear. I can recall many parents who look relieved to see that the urine looks negative and will more easily accept that the fever is in fact due to a viral infection.

    Also, we have seen many patients in the urgent care because they saw their doctor or went to the ER and was sent home with the diagnosis of a viral infection but “no testing was done.” We may feel justified in our diagnosis and think we saved a few bucks but the patient walked away not convinced when they didn’t see the evidence; and ended up paying for another visit somewhere else. In the end, we didn’t cut costs because there are now 2 visits to pay for.

  4. An easy treatment for urinary tract infections and suitable for paediatric patients is D-Mannose, a member of the sugar family found in fruits such as peaches and cranberries. It prevents E.coli from adhering to the urinary tract walls. Usual clearance times for infections are less than 48 hours. If an infection is still present after this time suspect another micro-organism.

  5. According to these data, roughly 5% of children this age run around with pyelonephritis. Given that the average toddler has several (let’s say 3-5) febrile episodes, is it that the same 5% of children have multiple episodes, or, > 15% have at least one pyelonephritis a year?

Leave A Reply