Urinary tract infection is more common than once thought, and also can be somewhat subtle in its presentation. In some cases, children who seem to have simple upper respiratory tract infections or otitis media actually have an underlying urinary tract infection. UTIs are particularly common in females and uncircumcised males.
In part 2 of this series (EMN 2002;24:10), I discussed a unique study that attempted to prove that urine clarity alone could identify a urinary tract infection. Although it wasn't 100 percent accurate, the study found that clear urine was less likely to contain pathogenic bacteria. In last month's column, I discussed a study of urine obtained with a collection bag versus urine obtained with a urinary catheter. The results were unequivocal; urinary bags, though more convenient to use, are ineffective because the urine collected in them is often contaminated.
Can urine collected by catheter ever be contaminated? In theory, it can, and many authorities recommend discarding the first few drops or milliliters and actually sending a midstream specimen for analysis. But is this really necessary? Even a few drops of urine can be difficult to obtain in some children, and discarding these drops seems to make little sense provided sterile preparation was adequate.
Early vs. Later Stream Urine
In a study published in 2000, investigators attempted to determine whether later stream urine was more sterile than early stream urine. These authors obtained two urine samples from 86 children. The first three ccs, which represented their laboratory's minimum requirement for a complete urinalysis, gram stain, and culture, was called early stream urine and the remainder of the urine obtained was termed later stream urine.
A standard preparation technique (three Betadine soaked swabs) was used for each patient, and a catheter then was inserted into the urethra. After the early stream sample was collected, a second collection tube was used to collect the later stream sample. The catheter was not changed between samples. The determination of a true urinary tract infection was made using a comparison of the culture results of the two samples. If pathogenic bacteria were grown from both specimens, the patient was deemed to have a urinary tract infection. Conversely, if the culture from the early stream sample grew bacteria but the later stream urine did not, the result from the early stream urine was determined to be a contaminant.
Eighty-seven patients were enrolled in the study, but one was eliminated due problems with data collection. Most of the patients were quite young. They had a mean age of 250 days, and the median age was 98 days. Thirty of the patients were under 3 months old. Fifty-five of the patients were male. Only six patients were determined to have a true urinary tract infection.
In all six cases, the cultures from both the early stream urine and the later urine grew a single pathogenic organism. In all six cases, bacteria were seen on gram stain of both the early and later stream urine. In four of the six early stream specimens and in all six of the later stream specimens, bacteria also were seen on the urinalysis. There were, however, some significant differences between the two samples and most of these were seen in the analysis of the urine samples that did not grow pathogenic bacteria. In 16 of the early stream samples from these patients, bacteria were seen on the urinalysis, but when the later stream samples were analyzed, bacteria were only seen in 10 samples.
Likewise, 20 early stream samples but only 10 later stream samples contained greater than or equal to 5 white blood cells per high-powered field. Additionally, four of the cultures from the early stream urine grew bacteria. In all cases, the colony counts were quite low, and one of the organisms proved to be coagulase-negative staphylococcus, an unlikely cause of urinary tract infection.
These results are certainly not dramatic. In a situation in which one could only obtain a few drops of urine from a febrile infant, there seems to be little harm in sending the sample for culture. After all, as these authors point out, given the low colony counts, no one is likely to confuse the contaminated specimens with a true urinary tract infection. This is somewhat reassuring because although we don't want to miss a true urinary tract infection, neither do we want to subject a child to a needless and potentially invasive evaluation.
On the other hand, a significant number of samples of early stream urine contained five or more white blood cells, and in half of the cases these high white blood cell counts were not seen in the later samples. To a lesser extent, the same was true for bacteria seen on the urinalysis. Under real conditions, many of these infants and children would have received an unnecessary course of antimicrobial treatment. That certainly would not have been a terrible tragedy, but it is something that we'd like to avoid if possible.
How should the results of these studies influence practice? For starters, we have to be somewhat vigilant in our attempts to identify urinary tract infections. More importantly, we need to discard our urine collection bags and stick with catheters. When the fates provide us with an ample supply of urine, we should at least consider discarding the first one to three ccs and analyze a mid-stream sample. I recognize that this recommendation is not always practical. In many emergency departments, pediatric urine collection is performed using a self-contained kit that comes with a collection tube and a bladder catheter prepared as one unit. But based on the results of this study, it seems that we should try to use a mid-stream sample whenever possible.