A 4-month evaluation of ambulatory patients with a suspicion of a urinary tract infection was performed. Specific objectives included assessment of five urinary screening methods, reevaluation of the necessity of the phenylethyl alcohol plate (PEA), and cost-effectiveness of screening for low colony count bacteriuria. Urine samples were collected as midstream, clean-caught specimens. A total of 142 samples, 87 from 79 symptomatic patients and 55 negative controls, were evaluated. All urine specimens were cultured using a 0.01 mL loop and a 0.001 mL loop onto Columbia sheep blood agar, MacConkey agar, and PEA agar. Twenty-four specimens (17%) were sterile, 64 (45%) were contaminated, and 54 (38%) were infected. Five urine screening methods were performed. These tests and their associated sensitivity and specificity are as follows. The Chemstrip 9 (Behring, Inc., Somerville, NJ) for leukocyte esterase and nitrate, 67%, 98%; microscopic analysis on spun urine, 79%, 93%; methylene blue stain for pyuria, 60%, 99%; Gram stain for pyuria, 45%, 93%; Gram stain for bacteriuria, 65%, 75%; and the URISCREEN (Analytab Products, Plainview, NY), 92%, 89%. Inclusion of a PEA plate for isolation of gram-positive organisms provided no additional information. Routine culture of urine samples at 10(-2) mL increased the contamination rate by 19%.
