Abstract
The 1-min leukocyte esterase (LE)-nitrite test (Chemstrip 9; Biodynamics, Division of Boehringer Mannheim Biochemicals, Indianapolis, Ind.) and a bioluminescence assay (Monolight centrifugation method; Analytical Luminescence Laboratory, Inc., San Diego, Calif.) were tested for their efficacy as urine screens among 453 patients at a tertiary-care teaching hospital. Both methods had the capacity to exclude significant bacteriuria (greater than or equal to 10(5) CFU/ml) when compared with the results of conventional culture methods, with predictive values of 99 and 93%, respectively, for a negative test. Bioluminescence was the more accurate nonculture method used. Sensitivity and specificity values were 97 and 71%, respectively, for bioluminescence, 82 and 60%, respectively, for LE with nitrite, and 72 and 64%, respectively, for LE without nitrite. At reduced levels of bacteriuria less than 10(5) CFU/ml), the sensitivities of LE-nitrite and bioluminescence were decreased but comparable. The addition of protein and blood test results in the Chemstrip 9, along with LE-nitrite as bacteriuria indicators, were unsatisfactory because of the large numbers of false-positive results attributed to protein and blood determinations. LE activity as detected by the LE test was a poor predictor of significant bacteriuria in both male and female patients. The sensitivity (71%) and specificity (57%) of the LE test in male patients were significantly lower than those previously reported and varied with the patient population studied.