It is clear from recent studies highlighted in Koyle and Shifrin’s review that the past decade has seen a remarkable retreat from previous dogmas regarding UTIs. There are no good data demonstrating a relationship between febrile UTIs and subsequent renal diseases, and if there is such a relationship, it must be weak. Although we know that vesicoureteral reflux (VUR) is frequently found in children with a history of febrile UTIs, we admit that most VUR resolves spontaneously and that we do not have evidence as to whether treatment of the rest improves outcome.

These changes in the outlook toward UTIs are reflected in new, more conservative imaging recommendations, but their implications have not yet been sufficiently appreciated for the decisions pediatricians face most often when evaluating a young child with a fever: whether to obtain a urine sample, and if so, how to do so.

The decision to obtain a urine sample for testing should be based on both the probability of UTI and the projected health benefit of diagnosing it at the current visit. The former can be estimated by considering clinical and demographic risk factors, including absence of another likely source for the fever, fever duration, race, sex, and circumcision status. The health benefit of early diagnosis of UTI is the improvement in health outcomes that can be accomplished by treatment. Now that we acknowledge that we do not know the extent to which imaging or treatment of UTI actually prevents late adverse outcomes, we can focus on short-term outcomes, such as reduction of symptoms and prevention of immediate complications, such as sepsis or meningitis.

Because febrile UTIs are easily treated, the benefit of reducing symptoms is directly related to symptom severity. Patients who present with high temperature, flank pain, and longer duration of illness (suggesting that their UTI is not resolving spontaneously) have the most to gain from having their UTIs diagnosed and treated sooner. Serious complications of UTI, such as bacteremia and meningitis, are rare and largely seen in infants younger than 6 months, especially those younger than 2 months. It makes sense to test more aggressively for UTI in these infants.

What happens to children in whom a UTI is initially missed? In a prospective study of 15,781 episodes of febrile illness in children younger than 5 years, even those children whose bacterial infections were not initially diagnosed and treated did well. Data from the Pediatric Research in Office Settings (“PROS”) Febrile Infant Study suggest that even in the highest risk 0- to 3-month age group (in whom I recommend urine testing), the outcome is generally benign. In this study, there were 807 infants aged 0 to 3 months with a temperature greater than or equal to 38°C, whose practitioners did not initially order for any urine tests, who were not initially treated with antibiotics, and were followed up until resolution of their illness. Based on the demographic and clinical risk factors for UTI recorded at their initial visit, about 61 patients (7.6%) would have had a UTI diagnosed if a urine test had been ordered. However, only 2 patients were diagnosed with UTI; both were treated and did well. The other approximately 59 infants (97%) recovered uneventfully, without ever having their UTIs diagnosed. Given this benign outcome in a group at highest risk of complications, it seems that the short-term risk of failing to diagnose UTI at the initial clinic visit is extremely low. Hence, the benefit of early diagnosis of UTI is primarily the opportunity to reduce symptoms sooner and prevent a small number of return visits for persistent fever.

An open question is at what risk of UTI these modest benefits justify the cost, discomfort, and inconvenience of urine testing, as well as the accompanying risk of false-positive results. The American Academy of Pediatrics practice guideline suggests that this point is 1% to 2%, implying that it is worth performing urine tests on 50 to 100 children to identify a UTI (sooner) in 1 child. This seems like a very low threshold (and high number needed to test) to me. However, this threshold varies with the treatment and risk preferences of the family and clinician, as suggested by Drs Koyle and Shifrin.

Families and clinicians will also have a range of preferences regarding the relative value of a more accurate urine culture result obtainable by urethral catheterization weighed against the invasiveness of the procedure and the risk of introducing a UTI. Greater certainty is required if the consequence of a positive culture is to be imaging, particularly with a voiding cystourethrogram (VCUG). On the other hand, if the main decision made as a result of the urine culture is whether to discontinue antibiotics (already started because of a positive urinalysis result), it may not be worth doing about 20 urethral catheterizations to avoid administering antibiotics for an additional 8 days in 1 child.