Antimicrobial sensitivity results from a previous infection are a useful guide to empirical treatment of re-infection, a study from Ireland has shown

The researchers aimed to analyse sensitivity patterns to identify predictive values of previous E. coli isolates for the treatment of re-infections in clinical practice.

GP records of patients with more than one urine sample with significant bacteriuria were identified from a database. Recurrent infection was defined as re-infection with the same bacterial species after a period of more than two weeks. Re-infection within two weeks was considered to be relapsing infection, rather than recurrence, and these cases were excluded.

Overall 147,306 urine samples were analysed. In 21.3% of samples a causative organism was identified, this was found to be E. coli in 14.4% of cases. A total of 3,413 patients were identified with more than one E. coli positive sample over the 4.5 year study period. Of these, 1,092 had had recurrent infection within three months, 693 between three and six months and the remainder between six months and one year.

The prevalence of resistance to common antimicrobials was: 2.6% for nitrofurantoin, 26.4% for trimethoprim and 60.7% for amoxicillin. Studies have suggested that an antibiotic is unsuitable for empirical use when resistance is greater than 20%, which would indicate that, in this community at least, trimethoprim is no longer suitable as a first-line empirical treatment.

There was an 85% probability that a patient with an ampicillin-resistant E. coli in a previous UTI would still have an ampicillin-resistant E. coli in a further episode of bacteriuria within three months. Respective positive predictive values were 84% for ciprofloxacin, 78% for trimethoprim and 20% for nitrofurantoin.

The probability that a re-infection within three months and up to a year is susceptible to a particular antibiotic if the initial E. coli was susceptible to that drug was nearly 100% for ciprofloxacin and nitrofurantoin, but 86% for trimethoprim.

The authors conclude, therefore, that if resistance to amoxicillin, trimethoprim, or ciprofloxacin is detected, a further prescription for recurrent infection within three months is not advisable, while susceptibility to nitrofurantoin, ciprofloxacin, or trimethoprim in a previous sample supports the prescription of these drugs for a re-infection within three months and up to a year.

Resistance to nitrofurantoin was low in this sample, and even when detected was found to decay relatively quickly. This study would suggest that nitrofurantoin is now, therefore, a better choice of empirical antibiotic for UTI unless previous MSU results have suggested trimethoprim susceptibility (within the previous year). This is consistent with the latest guidelines from the Health Protection Agency (HPA), recommending the use of nitrofurantoin as first-line empirical antibiotic in UTI.

Dr Jonathan Rees