Urinary tract infections

April 2004

---Philip A. Brunell, MD

What do you do when you have a febrile infant from whom you have obtained a bagged urine, which is now growing innumerable organisms, who has already been started on antibiotics? Unfortunately, this is one of the more frequent and frustrating calls received by infectious disease consultants. The problems that arise in this situation are not only the immediate treatment plan but, at least as important, whether to proceed with additional work-up and what sort of follow-up is indicated. The scenario described is made even more complicated when this laboratory datum is recorded on the child’s chart, especially in the absence of a note indicating how the urine was collected. Making this a part of the child’s permanent medical record will impact her or him, including the management of future febrile episodes. This is not a nice thing to do to a patient.

Unfortunately, we showed a child with an attached bag for collection of urine with an article summarizing Dr. Ellen R. Wald’s presentation at a recent AAP meeting (“Enhanced urinalysis boasts optimal rates for finding UTI,” February.) “An enhanced urinalysis [my italics] is performed on all emergency-room patients who have a catheterized urine obtained, usually for the evaluation of fever,” Wald stated in the article.

In the practice guidelines of the AAP on urinary tract infections (UTI) (Pediatrics. 1999;103:843), it is stated that “a urine specimen should be obtained by suprapubic aspiration [SPA] or transurethral bladder catheterization; the diagnosis of UTI cannot be established by a culture of urine collected in a bag (strength of evidence: good).” They go on to state that with a prevalence rate of true UTI of about 5% and a false-positive rate of about 70%, which can be expected from a bagged specimen, bagged specimens with a significant colony count would be falsely positive 85% of the time. To quote Dr. Doug Baker, speaking at the 16th Annual Infectious Diseases in Children Symposium in New York, “The only reliable methods of obtaining uncontaminated urine samples from infants are needle aspiration or catheterization of the urinary bladder. Specimens collected from externally applied bags are worthless from the perspective of evaluation of fever in these infants.”

In the AAP guidelines there is an alternative suggested for certain situations where treatment can be withheld and, presumably, there is a strong family objection to SPA or catheterization (Pediatrics. 1999;103:843). If this is done, the parent must be apprised of the need to withhold antibiotics until a proper specimen is obtained and the consequences, described above, of treating solely on the basis of a bagged specimen. Although there is a dearth of randomized controlled trials on delaying treatment as opposed to empirically starting treatment, at least one study showed that delaying treatment will not affect the frequency of scarring at six months post infection (N Engl J Med. 2003;348:195-202).

It is suggested in the guidelines that if the culture of urine obtained from a bagged urine is negative, infection can be ruled out and more invasive methods of obtaining urine can be avoided. Surrogates of a significant colony count (eg, determination of urinary nitrates or leukocyte esterase), tests which were designed for use in adults, are very insensitive in children.

More than 40 years ago, Dr. Stansfeld, an English pediatrician, showed the relationship between counting white blood cells (WBC) in uncentrifuged urine and finding more than 100,000 colonies of bacteria in the urine specimens of children (Arch Dis Child. 1962;37:257). Dr. Hoberman has reported that the presence of >10 WBC/mm³ in uncentrifuged urine obtained appropriately and not in a bagged specimen correlates with a colony count of >50,000 in 84.5% of specimens. Routine urinalysis in which urine is centrifuged and WBC per high-power field enumerated in only 61%. Dr. Pryles found a somewhat lower sensitivity, but he used a criteria of >100,000 colonies on “clean urines” (Am J Dis Child. 1965;110:628). Some question the need to obtain a urine culture on infants who are febrile without an obvious source, but it is estimated that about 5% will have UTIs. The AAP guidelines indicate that renal scarring is more likely to follow UTIs in young infants, and repeated infections increase the risk (Pediatrics. 1999;103:843). Thus, it is important to diagnose UTI in these infants, both for treating the initial infection as well as for preventing subsequent ones. Although most of these fevers in infants will have causes other than UTI, there is no way of ruling them out other than by obtaining a negative culture from an appropriate urine specimen or by signs suggesting another etiology.

Scanning options

Prevention of renal scarring and relief of acute symptoms are the two goals of treating UTIs. Intuitively, prevention of scarring would appear to be desirable, but there is a dearth of good data on the impact of renal scarring on renal function and hypertension in later life. Some of the estimates of renal scarring were made on the basis of intravenous pyelogram (IVP) or ultrasound examination, both of which are less sensitive than the Technetium-99m–labeled dimercaptosuccinic acid (DMSA) renal scan. This is the most reliable method of evaluating renal scarring and diagnosing pyelonephritis as distinct from cystitis. The exact role of DMSA in management of these infants is still being debated. In a recent report, it was shown that only 9.5% of children with first-time UTIs had renal scarring six months post onset, and these involved an average of only 8.2% of renal tissue (N Engl J Med. 2003;348:195-202).

DMSA scanning in the acute stage of first UTIs can separate pyelonephritis, present in 61%, from cystitis, but as the treatment plan will not be altered and residual scarring cannot be predicted from these studies, DMSA scanning is not recommended at this time. Voiding cystourethrogram (VCUG) should be performed at this time (N Engl J Med. 2003;348:195-202) as 39% of infants with proved UTI were found to have reflux on VCUG; only 4% were greater than grade III. This is useful as a routine to detect anatomic abnormalities and to establish whether reflux is present. Although spontaneous resolution of ureteral dilation observed in utero does not appear to result in permanent renal damage (J Urol. 2001;165:2000), the combination of infection and reflux may. The presence of reflux during acute infection was associated with residual renal scarring (P=.03), the frequency being related to the degree of reflux (P=.0007) (N Engl J Med. 2003;348:195-202). Prophylaxis is used during the period of diagnosis and VCUG and in patients with . grade I to prevent infected urine from refluxing into the kidney, anticipating that further damage will be prevented.

Ultrasonography was found to be insensitive, detecting only higher grades of reflux. Moreover, routine examination by ultrasonography in children failed to detect appreciable numbers of clinically significant abnormalities (N Engl J Med. 2003;348:195-202). However, it is likely that many of the infants in this study had already had been studied in utero. Routine fetal ultrasonography has reduced the usefulness of this procedure postnatally in evaluation of infants with UTI. It is estimated that about 2% of infants have some degree of hydronephrosis in utero, half of whom resolve spontaneously at birth and the additional number by the first birthday (J Urol. 2001;165:2006). Similarly, reflux is much more common at birth and tends to diminish with maturity during childhood. Ultrasonography may have a role in UTI not responding to therapy and might be considered for detecting abnormalities if prenatal ultrasonography has not been done.

Clean catch

Less invasive methods of collecting urine may be appropriate in older children who can voluntarily provide a urine sample. Most of the data on “clean catch” are based on urines collected by trained nurses using strict guidelines for preparation of patients and for urine collection and not by parents at home. Urine specimens were cultured promptly or refrigerated. One urine specimen containing >100,000 colonies/ml obtained by this method will be adequate to make the diagnosis in 80% of cases. Increasing the number of specimens increases the likelihood that a UTI is truly present (Pediatrics. 1999;103:843).

Infants who are febrile and do not have an obvious source have about a one in 20 chance of having a UTI. There is no way of selecting the affected infants other than doing a properly collected urine culture. As these infants appear to have a greater risk of renal scarring than older children, and scarring appears to be more likely with repeated infections, identification of these infections is important. Unfortunately, the long-term risk of scarring has not been adequately evaluated, although intuitively it would seem that it would be a good thing to prevent. There is virtually unanimous opinion that bagged urines should not be used to determine whether these febrile infants have a UTI. Catheterized specimens or those obtained by suprapubic aspiration are required. Surrogates, eg urinalysis, including detection of urinary nitrates or leukocyte esterase or cell counts, cannot substitute for a properly done culture. Labeling a child as having a UTI on the basis of improper criteria is a disservice to these patients.