OBJECTIVE: In a previous study, elevated urine antibody levels in institutionalized subjects with asymptomatic bacteriuria were associated with decreased survival. This study was undertaken to confirm the previous observation in a prospective study in a larger cohort and to explore selected other variables that may be associated with survival. DESIGN: A prospective, 24-month, observational cohort study. SETTING: Three large nursing homes in Winnipeg, Manitoba. PARTICIPANTS: Permanent residents were identified by initial screening urine cultures and subjects with bacteriuria were enrolled. The median age of subjects was 76 years, 51% were women, and the median duration of residence before enrollment was 26 months. Subjects were highly functionally impaired. MEASUREMENTS: Monthly urine specimens were collected for culture, leukocyte count, and urine antibody. Serum specimens for antibody to uropathogens and IL6 were obtained at enrollment and every 6 months. Anthroporphometric tests of nutritional status and functional and mental status were also measured every 6 months. Residents were stratified as having elevated or not elevated urine antibody, based on the initial urine specimen. The mean urine antibody for all of each subject's specimens was also calculated, and subjects were stratified as low, intermediate, or elevated urine antibody. Outcomes measured included mortality, infection and antibiotic use, and functional, mental, and nutritional decline. RESULTS: Ninety-eight bacteriuric subjects were enrolled in the study; 34 (35%) had elevated urine antibody and 64 (65%) had not elevated urine antibody. The two groups did not differ in demographic features, co-morbidities, functional status, medication use, or infecting organisms. Survival was significantly (P <.001) poorer in the group with elevated urine antibody. At 24 months, 35% of subjects with an elevated urine antibody were alive compared with 75% with a low urine antibody level. This survival difference was consistent when the two groups were stratified by sex, institution, and presence of a chronic indwelling catheter. Subjects with elevated urine antibody had no evidence for accelerated functional or nutritional decline during the study period compared with the group with low urine antibody. These subjects did, however, have an increased incidence of episodes of symptomatic urinary infection and infections at non-urinary sites. CONCLUSIONS: Older, bacteriuric, institutionalized subjects with elevated urine antibody had decreased survival rates compared with subjects with lower urine antibody levels. There is no clinical evidence to support accelerated decline caused by chronic infection to explain this observation. Urine antibody may be a marker for immune dysregulation, which precedes death in older impaired subjects. J Am Geriatr Soc 46:947–953, 1998.