Antibiotic resistance in uncomplicated urinary tract infection: problems with interpreting cumulative resistance rates from local community laboratories.

AIMS: To determine the resistance rates and patterns in bacteria causing uncomplicated urinary tract infections (UTIs) presenting to general practitioners (GPs) in Christchurch. METHODS: 82 randomly selected GPs in Christchurch participated in the study. Midstream urine (MSU) samples were prospectively collected for standard microbiological analysis on all women between the ages of 16 and 50 years presenting with symptoms of dysuria and frequency and who had positive dipstick testing to either (or both) nitrites or leucocytes. MSUs were submitted for bacterial colony counts and resistance testing of isolates present in adequate numbers. RESULTS: 374 specimens were collected. 299 filled the inclusion criteria, of which 94 fulfilled criteria for significant infection. Trimethoprim resistance was found in 8, (8.5%) (95%CI 2.8,14.2) overall with a resistance rate for Escberichia coli (E. coli) to trimethoprim of 11.5%. This compared with cumulative resistance rates from local community laboratories for E. coli to trimethoprim of 19%. For a woman in this age group presenting with symptoms of UTI we estimated that her probability of having a trimethoprim resistant organism was 2.7%. CONCLUSION: Trimethoprim remains a reasonable first line treatment for uncomplicated UTI in Christchurch. Actual resistance rates are significantly less than those derived from routine pooled laboratory specimens, and when used in an intention to treat calculation to inform empiric prescribing, become even less significant. While collection of these routine data is essential to provide early warning of emergent resistance, a truly representative rate should be determined to inform prescribing decisions if resistance appears to be increasing.