a years ago when I was a medical student, the instruc-sever tor for I tlie renal physiology lecture began his reniarks with the words, "You are what you urinate." It was not until I began the clinical practice and study of internal medicine did I realize the true meaning and relative validity of that prophetic comment. Certainly, a careful exaniination and analysis of urine will often provide a vast amount of information about patients, including normal physiology, changing homeostasis, and underlying diseases or disorders. Moreover, findings in the urine niay have predictive value: early detection of proteinuria in diabetics heralds the onset of diabetic ~iephropathy.'-~ Earlier studies of institutionalized older residents indicated that the presence of bacteriuria (generally asymptomatic) predicted a higher mortality rate cornpared with residents who were aba~teriuric?-~ I-Iowcver, subsequent investigations refuted these findings and suggested no such association between bacteriuria and niortality.8-'1 Furthermore, when antimicrobial therapy was administered to bacteriuric (symptomatic and asymptomatic) institutionalized older adults, mortality was not substantially affect-So what does the clinician conclude from these conflicting findings?To add further to the challenge of better understanding tlie relevance of asymptomatic bacteriuria in this population, Nicolle and colleagues, in this issue of JAGS," have investigated what might be factors that influence mortality in institutionalized residents who are bacteriuric. In an earlier preliminary study, Nicolle and colleagues found that institutionalized bacteriuric older adults with elevated levels of urine antibody to tlie major outer membrane of Escherichia coli had a survival rate of approximately 20% of those subjects with normal urine antibody levels at the end of 3 years.'6 Moreover, bacteriuric older people with normal urine antibody levels had survival rates similar to abacteriuric residents. In the current study," this observation has been pursued further by a prospective, 24-nlonth observational cohort investigation of 98 bacteriuric older residents (age range 61-94 years) from three long-term care facilities. Again, it was found that there was a correlation between high levels of urine antibody (to outer membrane antigens of E. coli) and increased mortality (68% vs 17%) in comparison with study subjects who had bacteriuria and low levels of urine antibodies. This association remained valid regardless of gender, presence or absence of a urinary catheter, presence or absence of urinary symptoms, and study site. Curiously, in comparison with subjects with low levels of urine antibodies, the high-level group did not have greater or more rapid decrements in nutritional anthromorphonietric measures, physical functional capacity, or cognitive parameters. However, not mentioned in this study were the data on serum albumin, cutaneous anergy status, details of physical performance measures, relationship of urine antibody levels and ,d.12-14 elimination or persis...