Epidemiologic and diagnostic aspects of bacteriuria: a longitudinal study in older women.

Abstract

OBJECTIVE: To examine month-by-month variability of bacteriuria in a cohort of older women and to evaluate the performance of rapid diagnostic tests commonly used to indicate the presence of significant bacteriuria. DESIGN: Prospective, observational study. SETTING: Community housing sites and a long-term care institution. PARTICIPANTS: Sixty-one women, mean age 77.6, took part in the study. MEASUREMENTS: Midstream clean-catch urine samples and medical information on subjects were collected at baseline, and then monthly for 6 months. RESULTS: Bacteriuria alone (> or = 10(5) organisms per mL) occurred in 17% of all urine samples (28% of patients), bacteriuria with pyuria in 15% (26% of patients), and bacteriuria with symptoms in 3% (10% of patients). Spontaneous clearance of bacteriuria with pyuria was common (P = .30), as were new occurrences (P = .12) over 6 months of follow-up. For the outcome of bacteriuria with symptoms, sensitivity of urinary diagnostic tests such as bacteria and pyuria on microscopic analysis, and leukocyte esterase on dipstick testing, ranged from 79 to 93%. Negative predictive values of these tests approached 100%. CONCLUSIONS: Bacteriuria was a very common event, occurring in almost one-fifth of all urine samples and one-third of all subjects during 6 months of follow-up. Month-by-month follow-up indicates that the natural history of bacteriuria is marked by frequent spontaneous alternation between positive and negative events. The high negative predictive value of many simple diagnostic tests commonly used for urinary tract disease suggests that they can quickly and cost-effectively rule out bacteriuria in the older female patient.