Should We Abandon Standard Microscopy When Screening for Urinary Tract Infections in Young Children?

Dipstick analysis is a screening test for urinary tract infection (UTI) in office practice. Standard urine microscopy is available offsite. We evaluated the usefulness of these tests on consecutive urine samples obtained by bladder catheterization in young children with a possible UTI. Thirty of 230 children had positive cultures. Dipstick analysis was specific (98%) but not sensitive (70%). Standard microscopy was not specific. Dipstick analysis and urine culture are useful tests for the diagnosis of UTI. The diagnosis of urinary tract infection (UTI) is based on urine culture, 1 but the culture results are not available until 24 to 48 h after the patient provides the specimen. Thus several screening methods for evaluating a young child with fever and possible UTI have been developed. These methods include microscopic examination of the urine sediment (a time-honored method) and dipstick analysis for nitrite or leukocyte esterase. To increase the reliability of microscopy, an enhanced urinalysis has been studied for diagnosing UTI in the pediatric population. In this method uncentrifuged urine is Gram-stained for bacteria, and a leukocyte count is performed using a hemocytometer. Several studies 2–5 have demonstrated that urine Gram stain has increased sensitivity and specificity compared with standard microscopy for predicting urine culture results. A recent American Academy of Pediatrics (AAP) practice parameter 6 endorsed urinalysis (UA) as a screening method for UTI in young children and stated that any of the following results on a UA may be suggestive of a UTI in a young child or infant: positive leukocyte esterase or nitrite on dipstick analysis; >5 white blood cells per high power field (hpf) on spun urine sediment; or the presence of bacteria in an unspun, Gram-stained specimen. In physician office laboratories Gram stain and microscopic examination of the urine are not generally available because these tests are not waived under the Clinical Laboratory Improvement Amendment of 1988 and require laboratory certification from the Health Care Financing Administration (HCFA). Of the 95 088 physician office laboratories in the United States, only 24% have a certificate of compliance or accreditation necessary to perform Gram stain and/or microscopy. An additional 32% have a certificate of provider performed microscopy that permits examination of urine sediment (personal communications, February 2001; Division of Laboratories and Acute Care Services, Survey, and Certification Group, HCFA). The actual percentage is probably much lower; however, numbers on individual tests are not kept by HCFA. Thus most physicians in office practice must rely on the dipstick analysis, which is a waived test under the Clinical Laboratory Improvement Amendment and an offsite laboratory for microscopy when evaluating a young child for a possible UTI. In our area of central Virginia (one university hospital, one community hospital and two private pediatric practices), the certified laboratories offer standard microscopy and not Gram stain. As previously noted 7 the enhanced UA is not used routinely in most hospitals. Because Gram stain of unspun urine is not usually available to pediatric ambulatory practice, we wished to evaluate the reliability of the UA available at our university hospital laboratory under nonstudy conditions. We retrieved data on the results of dipstick and standard microscopy on urine specimens from young children and infants for whom urine cultures were performed. Only samples obtained by bladder catheterization were studied to minimize inaccuracy in interpretation of both UA and culture. We particularly wished to ascertain whether standard microscopy offered any advantage over dipstick in evaluating young children for possible UTI. Methods. Design. Using the hospital database we retrospectively identified children <4 years of age seen as outpatients in the General Pediatric Clinic or Emergency Department at the University of Virginia Health Sciences Center. Subjects were included in the analysis if a urine sample had been obtained by in-and-out catheterization and a UA and culture had been performed. Data were collected from patients seen from January, 1995, to March, 1996, and from June, 1997, to May, 1998. This study was approved by the University of Virginia Human Investigation Committee. Laboratory evaluation. In general, urine samples were transported to the hospital clinical laboratory within 30 min after collection. Standard microscopy and dipstick analysis (Ames, Elkhart, IN, or Boehringer Mannheim, Indianapolis, IN) were performed. For standard microscopy urine was centrifuged for 6 min at 2300 rpm. The sediment was resuspended in a drop of supernatant and examined either manually or with an automated method 8 (The Yellow IRIS 900; International Remote Imaging Systems, Chatsworth, CA) for bacteria and white blood cells with a light microscope (×40); an average count was made of 10 hpf. Cultures were performed on 5% sheep blood-Trypticase agar and MacConkey agar (BD Biosciences, Rockville, MD). Organisms were speciated by an automated method (Vitek; BioMérieux, Hazelwood, MO). Definitions. Any amount of leukocyte esterase (including trace) or nitrite on the dipstick analysis was defined as positive. Bacteriuria by microscopy was defined as the presence of any bacteria and pyuria as the presence of ≥5 white blood cells/hpf. A positive bacterial culture was defined as growth of ≥10 000 colony-forming units (cfu)/ml of a uropathogen. Results. Thirty (12%) of the 260 patients reviewed had a positive culture. Escherichia coli was identified in 22 (73.3%) of the 30 positive specimens, Klebsiella pneumoniae in 3, Enterococcus faecalis in 3 and Enterobacter cloacae and Proteus mirabili s in 1 each. Twenty-one of the 30 cultures had growth of ≥100 000 cfu/ml, and 9 had ≥10 000 cfu/ml. One additional patient was excluded from analysis because of uncertainty about classification of the culture that grew >10 000 cfu/ml coagulase-negative staphylococci. Comparison of UA to culture results on the catheterized urine specimens in the 260 patients (Table 1) demonstrated that the dipstick analysis was 98% specific but was only 70% sensitive (9 of the 30 culture-positive urine samples were negative for both nitrite and leukocyte esterase). All five of the nitrite positive samples were also leukocyte esterase positive. Microscopy was slightly more sensitive for predicting positive culture results, but at the expense of a marked loss of specificity (Table 1). The presence of bacteria was reported in 24 of the 30 culture-positive specimens (80% sensitivity), but bacteriuria was also noted in 22% of culture-negative samples. The majority of the false positives had “few,” “occasional” or “rare” bacteria, but 9 (38%) of the 24 true positives had these same designations for bacteriuria. All 19 of the culture-positive samples with ≥5 white blood cells/hpf by microscopy were leukocyte esterase-positive; 21% of culture-negative urine samples also had ≥5 white blood cells/hpf by microscopy.Table 1: Urinalysis and culture results on 260 urine specimens obtained by bladder catheterization for evaluation of suspected UTI in children <4 years oldCombining dipstick analysis with microscopy increased sensitivity further at the expense of reducing specificity to 64%. Nine (30%) of the 30 patients with positive cultures had negative urine dipsticks. None of these 9 had pyuria by microscopy. Variable amounts of bacteriuria were present in the 9: none (2); rare (1); occasional (1); few (2); moderate (1); and many (2). Six of the 9 patients had cultures with ≥100 000 cfu/ml E. coli (4) or K. pneumoniae (2), and the other 3 had ≥10 000 cfu/ml E. coli. Discussion. Our study provides compelling evidence that the standard microscopy available in the ambulatory setting is an inaccurate laboratory test for evaluating urine in infants and young children with possible UTI. Although its sensitivity may be good, microscopic examination has a high false positive rate that might lead to overdiagnosis of UTI and unnecessary treatment. Furthermore our results and those of others 7, 9 demonstrate that this technique provides no advantage over dipstick analysis. The dipstick analysis, which is available to most practitioners because it is a waived laboratory test, is specific and can help guide in the management of children being evaluated for UTI. 7, 10 Nitrite, which is specific, is insensitive, presumably because a positive result requires the presence of bacteria in the bladder for several hours, which does not occur in this age group. The recent AAP practice parameter has recommended dipstick, standard microscopy or Gram stain as useful screens when evaluating young children for UTI. Although presented with good intentions, these recommendations for screening urine in young children do not represent what is available to most physicians in ambulatory settings. Gram stain of unspun urine is sensitive and specific, but it is not available or permitted in most ambulatory practice settings. Standard microscopy, which is available only onsite to approximately one-half of practitioners, is time-consuming and has poor specificity. Although Gram stain and microscopy might be available by processing urine samples through an offsite laboratory, logistically this could be difficult to accomplish expediently as the practice parameter recommends. As a result of these limitations, we recommend abandoning standard microscopy when evaluating febrile infants and young children for UTI. Acknowledgments. We thank Keith R Powell, M.D., Sharon Dudley, M.T. (ASCP) and Alzena Bibb for their contributions to this work. This work was supported in part by the Pendelton Pediatric Infectious Disease Laboratory.