Dr. Hyman questions the criteria we used to define a case of asymptomatic bacteriuria (two cultures with 2 weeks of the same organism at a titer of 10⁵ CFU/mL or more in an asymptomatic volunteer) [1]. He apparently also questions our view that screening and treating asymptomatic bacteriuria in elderly ambulatory women do not appear warranted [1].

His concerns are based in part on results from a computer model that simulates the turnover of bacteriuria in the urine under several conditions [2] and on nonquantitative descriptive studies of a new staining technique for visualizing bacteria in the urine [3]. The computer studies suggest that organisms with a long generation time will be eliminated from the urinary tract if no source for continued seeding of the urine exists. However, when such a source is present, the titers in the urine will remain low and will not reach levels conventionally considered significant for urinary tract infection. The new staining technique reportedly allows identification of organisms not seen with the usual Gram stain. Together, these studies are interpreted as suggesting that the use of conventional methods, as in our study, would miss low-level bacteriuria. Dr. Hyman believes that detection of these organisms might require use of "more generalized media" and asserts that treatment of bacteriuria so detected might be life-saving.

We agree that some patients with low-level bacteriuria are symptomatic and that low-level bacteriuria may exist in some patients who are asymptomatic. If treating asymptomatic patients who have titers of 10 (⁵) CFU/mL or greater is not advantageous, however, it is highly unlikely that detection and treatment of asymptomatic patients with low titers would be advantageous or "life-saving." We agree that symptomatic urinary tract infection should be treated regardless of titer.