To investigate whether 10 days of rifampin-trimethoprim (RIF-TMP) or 6 weeks of sulfamethoxazole-trimethoprim (SMX-TMP) would decrease the relapse rate in patients with acute uncomplicated upper urinary tract infections in comparison with 10 days of SMX-TMP, we randomized 189 patients to receive RIF-TMP or SMX-TMP in a ratio of 1:2. After the site of infection was established by the antibody-coated bacterium (ACB) test, patients with upper-tract infections who received SMX-TMP were again randomized and received either a total of 6 weeks or 10 days of therapy. All patients who received RIF-TMP were treated for 10 days. Clinical and microbiological evaluations were repeated at 2 and 6 weeks posttreatment. Eighty-five patients (54 ACB positive) received 10 days of RIF-TMP, 71 patients (45 ACB positive) received 10 days of SMX-TMP, and 18 patients (18 ACB positive) received 6 weeks of SMX-TMP. The overall recurrence rates in patients who received 10 days of therapy were 32% for RIF-TMP and 23% for SMX-TMP (P = 0.13). There were 12 (14%) relapses in the RIF-TMP group compared with 2 (3%) relapses in the SMX-TMP group (P = 0.01). In patients with upper-tract infections, the relapse rates were not statistically significantly different (P = 0.13). There were two (11%) recurrences (one relapse and one reinfection) in the 6-week treatment group. This 6% relapse rate was not different from the 4% relapse rate observed in patients with upper-tract infections who received 10 days of SMX-TMP. The number of patients who discontinued treatment because of an adverse effect in the 6-week SMX-TMP treatment group was significantly greater than those in the 10-day SMX-TMP treatment group (P=0.003) and the RIF-TMP treatment group (P=0.05). Ten days of SMX-TMP was as effective as 6 weeks of SMP-TMP or 10 days of RIF-TMP in the treatment of uncomplicated upper urinary tract infections and caused the fewest untoward effects.