Old tests improve diagnostic accuracy
It’s muggy outside and the Pediatric ED is as busy as ever. Winter virus season has given way to summer virus season in the never-ending circle of hot. So you’re really not surprised to pick up yet another chart that reads “fever” as the chief complaint. This one is a 14-month-old female with three days of fever, vomiting and decreased p.o. intake. The fever has been subjective, but the parents say she was “very hot.” She’s only vomited a few times and hasn’t had any diarrhea. The parents report she won’t eat anything and she’s making fewer wet diapers. No one else is sick in the home and she doesn’t attend daycare. Her past medical history is unremarkable and she’s received all her immunizations so far.
On examination she looks like someone who doesn’t feel well, but she’s not toxic-appearing. Her temperature is 38.5 (rectal) and she’s mildly tachycardic. Her mucous membranes are tacky, but her capillary refill is less than 2 seconds. Your careful examination reveals no source for the fever. You attribute her tachycardia to the combination of fever plus mild dehydration and contemplate your options.
Fever without a source in a fully-immunized, nontoxic 14 month-old does not require blood work these days. However, most experts agree that a female less than 24 months of age with no source and a temperature of 39 or greater should get her urine checked. This one doesn’t really meet criteria as her fever is less than 39 in the ED. However, her parents say she was very hot at home. OK fine – better safe than sorry. If you miss pyelonephritis, you risk renal scarring. You decide to go with the “very hot at home” rather than the measured temperature in the ED and order urine. Since she looks dry and is singularly unimpressed by your tempting array of clear liquids, you decide to give her some IV fluids while you await the results. You debate with yourself but decide to order bloodwork as well, since you’re violating venous space with your IV anyway. You write the orders and proceed down the hall to your next patient.
There’s been a recent shift at your institution in the approach to possible urinary infections in children. You probably are continuing to get catheterized specimens in younger children, allowing midstream clean catches in your potty-trained population. What’s new is an ‘enhanced’ urinalysis – a combination of a standard UA and a urine gram stain (the gram stain being done on unspun urine). The theory is that you will miss fewer UTIs that way – urinary dipsticks looking for leukocyte esterase and/or nitrites can miss up to 12% of UTIs in children. Some folks believe that’s because the urine in childrens’ bladders, when compared to adults, moves through too fast to trigger the chemical reactions that produce those by-products. UAs are helpful when they demonstrate a positive hemocytometer count (> 5-10 white cells per high power-field), but not all UTIs produce pyuria. The urine gram stain enhances the sensitivity of the urinalysis by detecting bacteria that have not triggered pyuria.
Time flies by and you bop in to reassess your little patient. She liked the fluids and antipyretics you gave her and she’s definitely more chipper than she was. She has finally agreed to a po challenge. Her bloodwork is not particularly helpful – white count of 10, normal chemistries. Her urinalysis is negative for leukocyte esterase, negative for nitrite and has only 1 white cell per hpf. In the old days, you would have called that negative for UTI and send her out with the diagnosis of viral syndrome. However, her urine gram stain comes back a few minutes later with > 10,000 bacteria so you give her a dose of ceftriaxone and send her home on cefixime for presumed pyelonephritis, with careful follow-up arranged with her primary care physician.
And two days later, your curiosity piqued by the positive gram stain in the face of what appeared to be a negative UA, you check the urine culture results. Sure enough – greater than 10,000 E.coli.
Since I saw this child, I have seen two other children who had no pyuria, negative leukocyte esterace and nitrite on UA but positive urine gram stains, both of whom went on to have positive urine cultures. Next time you get a cath urine on a child with a fever, consider adding a urine gram stain. It might improve your sensitivity.
Amy Levine, MD, is an assistant professor of pediatric EM at UNC Chapel Hill