Three techniques compared for detecting bacteriuria in symptomatic patients.

We wanted to determine whether the microscopic evaluation of urinary sediment could be replaced by either a biochemical determination (Chemstrip-9) or a colorimetric staining procedure (Bac-T-Screen), and to evaluate the feasibility of omitting from urinalyses attempts to culture urines. Cultures were considered positive when colony counts were greater than or equal to 10(3) for catheterized patients and greater than or equal to 10(4) for noncatheterized patients. The results of three separate studies on symptomatic patients showed a progressive decline in the sensitivity of the Chemstrip-9, which is a test for leukocyte esterase activity, and a difference in the sensitivity of the Bac-T-Screen between two of the studies. Neither test was consistently more sensitive or more predictive of a positive culture than was urine microscopy. By the end of the third study, we were convinced that the three methods are comparably sensitive and specific. Because 13 to 36% of positive cultures would be missed by these techniques, urine from symptomatic patients should routinely be cultured.