In the 6 years since Kass1firmly reintroduced the concept of counting bacteria in the urine,2physicians have become increasingly aware of the wide prevalence of urinary tract infections.1,3-5In addition, with the use of the quantitative urine bacterial count as an epidemiologic and experimental tool, many workers have contributed to the growing body of evidence implicating various modes of urinary tract instrumentation in the induction of bacteriuria, pyelonephritis, and their complications.3,6-14 The present study was conducted as a result of a systematic review of cases of bacteremia at the Jersey City Medical Center in 196015; this survey demonstrated a high incidence of bacteremia very similar to that reported by Finland16and revealed a close association between Gram-negative rod bacteremia and prior urinary tract instrumentation. Analysis of 81 cases of Gram-negative rod bacteremia complicating pyelonephritis (Table 1) indicated that in 78 the patient had