The efficiency of nitrofurantoin and trimethoprim prophylaxis in preventing recurrent urinary tract infections (UTI) was compared by means of actuarial percentage recurrence-free curves in a randomized, double blind study in 130 children (126 girls, 4 boys) aged 1 to 14 years (mean 7.5). The children received the antibiotics for 6 months. Nitrofurantoin proved to be the most efficient prophylactic drug in patients with abnormal urography and/or reflux (n = 60) as evaluated by actuarial percentage recurrence-free analysis (p = 0.0025). However, no differences was found in patients without urinary tract abnormalities. Nitrofurantoin prophylaxis altered neither the pattern of resistance nor the bacteriological constellation, while patients receiving trimethoprim prophylaxis had 76% trimethoprim resistant bacteria during prophylaxis, compared with 8% before (p less than 0.0001) and 17% after (p less than 0.0001) prophylaxis. The percentage of recurrences due to E. coli (70-80%) was unaffected by trimethoprim prophylaxis, but the proportion due to trimethoprim resistant E. coli was significantly higher during prophylaxis (65%) than before (6%, p less than 0.0001) and after (11%, p less than 0.001). The percentage of Staphylococcus epidermidis UTI was significantly higher during trimethoprim prophylaxis (27%) than before (2%, p less than 0.0003). Following prophylaxis there was no difference in the actuarial percentage recurrence-free curves of the two regimens. Side effects occurred more frequently in the nitrofurantoin group (37%) than in the trimethoprim group (21%) (p = 0.05). The majority of side effects in the nitrofurantoin group derived from gastrointestinal symptoms. In conclusion, nitrofurantoin is recommended as the first choice prophylactic treatment of children with recurrent UTI and urinary tract abnormalities.