There are currently no laboratory or clinical guidelines for the identification and treatment of disease caused by metronidazole-resistant strains of Trichomonas vaginalis. Fifty-three isolates of T. vaginalis from cases of refractory vaginitis in the United States (26 states) and Canada were tested for aerobic and anaerobic metronidazole susceptibility, and after various dosages of metronidazole, the therapeutic outcomes were evaluated for 31 of these cases. The mean aerobic metronidazole susceptibility of these isolates was 195.5 micrograms/ml (range, 12.5-greater than 1,000), which was about eightfold higher than that seen in isolates that were not resistant to metronidazole. The mean anaerobic susceptibility was 5 micrograms/ml (range, 1.6-25), which was about threefold higher than that of isolates from nonresistant strains. The average aerobic-to-anaerobic ratio of metronidazole susceptibility in the highly resistant isolates was more than 3.5-fold greater than that seen in the nonresistant isolates. White women accounted for 88% of the resistant infections. Of 31 cases that were re-treated and monitored, the highest average dose that failed to achieve a cure was 2.1 g of metronidazole/day given over an eight-day period; 27 (87%) of 31 cases were ultimately cured with an average dosage of 2.6 g of metronidazole/day given over a mean period of nine days. Resistance to treatment with metronidazole varied from mild to severe, and the resistance was occasionally more severe than the susceptibility values indicate.