Subtle changes in laboratory findings become more important when paired with this old standby
Business has been steady in the pediatric emergency department this afternoon. You’ve seen a wide range of cases today, which has been fortuitous as you have a medical student shadowing you who wants to “learn the ropes.” The nurses put a new patient in bed 2. You couldn’t ask for a better, basic teaching case. It’s a 4-month-old male with a fever. The baby looks great and is watching you with wide-eyed interest as you get the history from his mother.
You review his vitals and nothing stands out except a temperature of 38.1 rectally. He is cooing and interactive on the examining table and has no focal or concerning findings on physical exam. His anterior fontanelle is nice and soft. He is not circumsized. So, it is time to make some decisions. Should a nontoxic-appearing 4 month-old with no focus of infection evident on physical examination get a work-up? If so, what kind of tests would you get?
Much has been written about the febrile child less than 36 months of age. Everyone agrees that the child in this age range that appears toxic should be presumed to have a serious bacterial infection and managed accordingly. The discussion gets a little more lively when the infant looks well.
We decide that the infant is at some increased risk of a serious occult bacterial infection because of his young age and the fact that he has not had much in the way of immunizations. Because the infant looks great we decide that the likelihood of meningitis is low and we elect not to perform an LP. The other hospital did a chest xray yesterday and the child has no respiratory symptoms or findings on exam, so we choose not to repeat a film. We decide to get a CBC and blood culture to assess for occult bacteremia, which is rare in the child who has had multiple immunizations against s. pneumoniae and h.influenza type B, but might be present in this kid who has not had multiple rounds of vaccination yet. We also decide to check urine.
The CBC comes back with a white count of 21.7, but no left shift. The urinalysis, obtained by catheterization, comes back with 1+ leukocyte esterase, negative nitrites and only 2 white cells per high-powered field of uncentrifuged urine. “Ok, so it’s probably not his urine,” says the student. “Not so fast” you say, awaiting the gram stain result. Soon the lab reports gram negative rods in his unspun urine sample. You treat the child presumptively for pyelonephritis. Two days later, the blood culture remains negative but the urine culture grows e.coli.
Urinary tract infections can be difficult to diagnose in the diaper set. Unlike older children and adults, young children don’t come in with burning with urination, frequency, urgency or flank pain. Sometimes the only symptom they have is fever. With fever and a temperature of 39 or higher, the rate of urinary tract infection can be significant in children 36 months of age or less. There are reports that untreated pyelonephritis (assumed when the temperature is 39 or higher) can lead to hypertension or ESRD down the road. So it is important that the ED physician not miss this diagnosis. So how can you reduce the likelihood of a miss?
First, get a good specimen. No bagged urines, they get contaminated with the fecal flora resident in the diaper area no matter how well you clean your patient. Catheterization or supra-pubic bladder aspiration are the preferred methods for assessing the urine in a child too young to give a clean catch.
Second, don’t rely solely on the urinalysis. If it is positive, it is helpful but our patient had a very unexciting UA and we might have missed his infection if we had stopped with that. Looking at infants who are subsequently found to have a urinary infection by culture, up to 12% will have a negative UA. While nobody is sure why this is, there is the thought that nitrites and leukocyte esterase are by-products of the metabolism of bacteria and white cells and take some time to develop. Infants in diapers pee as soon as the urge strikes and may not retain the urine long enough to develop these by-products.
Third, get a gram stain. That’s what allowed us to pick up the infection in this child, even though the UA didn’t look like much.
Hoberman promoted the use of the “enhanced urinalysis” to avoid missing UTIs in infants. The enhanced UA is considered positive if either an unspun urine sample has greater than or equal to 10 WBCs/hpf on a hemocytometer cell count OR if the unspun gram stain shows bacteria. The enhanced UA is the most sensitive test to predict a positive urine culture but most labs don’t perform it. Our lab does the UA on unspun urine but does not perform the white count by hemocytometer because they claim it adds extra time. They do perform the gram stain on uncentrifuged urine, which is what allowed us to diagnose this child in the ED.
Fourth, send a culture and get a good phone number for the family so that therapy can be initiated or changed pending the results.
The point of this case is that this 4 month-old had pyelonephritis and we would have missed it if we had relied on the UA alone. The gram stain is what alerted us to the diagnosis. So, if you are going to get a urine specimen in a well-appearing child less than 36 months with a temperature of 39 or higher by history or in the ED, do this: cath specimen, UA AND gram stain AND culture. Initiate treatment in the ED if either the UA or the gram stain comes back positive.
A gram stain is great, add it to your list and avoid missing UTIs.
Hoberman, A., et al. Enhanced urinalysis as a screening test for urinary tract infection. Pediatrics 1993;91:1196.
Herr SM, et al. Enhanced urinalysis improves identification of febrile infants ages 60 days and younger at low risk for serious bacterial illness. Pediatrics 2001;108:866.
Amy Levine, MD, is an associate professor of pediatric emergency medicine at UNC Chapel Hill.