Comparison of long-term, low-dose pivmecillinam and nitrofurantoin in the control of recurrent urinary tract infection in children. An open, randomized, cross-over study.

Thirty-five children with a history of vesicoureteric reflux or with recurrent urinary tract infections were randomly allocated to low-dose prophylactic treatment with pivmecillinam or nitrofurantoin. After 6-10 months they were crossed over to the alternate drug for another 6 months, but only 24 completed the study because of lack of compliance or intolerance to nitrofurantoin. There was no significant difference in the long-term prophylactic effect between the two drugs, the overall infection rate being 0.7/patient-year. Pivmecillinam was significantly better tolerated than nitrofurantoin (P = 0.01). Nitrofurantoin effected no major change in the faecal flora, and nearly all urinary infections occurring during long-term treatment were caused by Escherichia coli. In contrast, a marked reduction of E. coli and a marked increase in Gram-positive cocci were found in the faecal flora during treatment with pivmecillinam. Seventy per cent of infections were caused by Streptococcus faecalis and only 20% by E. coli during pivmecillinam treatment (P = 0.001).