Aim-To compare the performance of leucocyte esterase and nitrite dipstick tests with microscopic examination and culture of first morning urines (n = 420) of hospital inpatients. Results-The sensitivity, specificity, and negative predictive value of the leucocyte esterase test for the cutoV of > 10 WBC/µl were 57%, 94%, and 68%, respectively. For > 5 WBC per high power field (HPF) these variables were 84%, 90%, and 93%. For > 10 5 colony counts/ml, the sensitivity of the nitrite test was 27%, specificity 94%, and negative predictive value 87%. When either leucocyte esterase or nitrite positivity was accepted as a marker of urinary tract infection, the sensitivity was 78%, specificity 75%, and negative predictive value 94%, and there were 22% false negative results. Semiquantitative microscopic estimation of bacteria per HPF yielded 40% false positives. Conclusions-Leucocyte esterase and nitrite dipstick tests are not suitable for screening for urinary tract infections. (J Clin Pathol 1998;51:471-472) Keywords: leucocyte esterase; nitrite; dipstick Screening of urine specimens for urinary tract infection by means of dipstick leucocyte esterase and nitrite test is not an uncommon practice, although the standard method for diagnosis of infected urine remains microscopic examination and quantitative culture of urine. We have evaluated the diagnostic performance of the dipstick test in hospital inpatients in comparison with the microscopic and culture results.
MethodsOver a three week period, 420 first morning urine samples from hospital inpatients on which urine cultures had been requested were randomly selected for additional tests. These tests included (1) test strip screening with multistix 8 SG (Bayer Corporation, New York, USA), (2) direct microscopic counting of the white blood cells (WBC) and bacteria per microlitre of urine using Kova counting chambers (Hycor Biomedical Inc, California, USA), (3) urine sediment microscopy, after centrifugation at 350 g for five minutes, to obtain WBC and bacteria per high power field (HPF). The strips were read by reflectance spectrophotometric method on Clinitek 200+ (Bayer).Urine cultures were performed by inoculating, from a standard loop, 10 µl of uncentrifuged and well mixed urines on to blood agar and MacConkey plates (Oxoid, Hampshire, UK) and incubating aerobically at 37°C overnight. Growth of < 10 000 colony forming units (CFU) per ml was considered negative.The patients consisted of 234 females (mean age 55 years, range 17 days to 97 years) and 186 males (mean age 58 years, range 9 to 94 years) and they belonged to the following general groups: paediatrics (4), oncology (5), obstetrics/gynaecology (39), surgery (51), geriatrics (57), ambulant consultation (68), internal medicine (93) and medical and postoperative intensive care (103).
ResultsUsing the standard criterion of > 10 5 CFU/ml for urinary tract infection, 72 patients' urines (17%) produced positive cultures. Of these 18 (25%) were infected with non-nitrate-reducing organism, these being mainly ...