Poster, Podium & Video Sessions
Presentation Authors: Michael Holmes*, Ray Littler, Megan Lyons, Lisa Smit, Glen Devcich, Adam Davies, John leyland, Chris Mansell, Hamilton, New Zealand
Introduction: Infective complications post transrectal ultrasound guided (TRUS) prostate biopsy appear to be increasing, probably linked to a rising prevalence of FRE in the bowel flora. Several authors have suggested potential patient factors which may increase the risk of FRE carriage or sepsis post TRUS prostate biopsy . National guidelines have suggested screening only high risk patients for FRE. We sought in a prospective study to assess the prevalence of FRE in our patients and whether previously identified patient factors were related to this. Can we identify a high risk group for FRE and disregard the rest?
Methods: A transrectal swab, screening for FRE, was taken prior to biopsy. Antibiotic prophylaxis was 1 gram p.o. of ciprofloxacin prior to and 500mg after biopsy. Targeted antibiotics were used if a FRE was identified. Information was collected on: 1. Previous number of TRUS prostate biopsies, 2. Overseas travel within the last 6 months and if to a developing country, 3. Diarrhoea while away, 4. Overseas travel at any stage and if to a developing country, 5. Antibiotic use within the last six months, 6. Diabetes and 7. Inflammatory bowel disease. Naive Bayes, Logistic Regression, and Random Forest classifiers were used to build predictive models. A leave-one-out validation was used to generate class probabilities to quantify expected performance.
Results: Rectal swabs were performed in 1135 of 1216 prostate biopsies. FRE was detected in 95 (8.4%) of which 16 were extended spectrum beta lactamase (ESBL) E.coli. The prevalence of patient risk factors are shown in Table 1. Travel to a developing country within 6 months, ever, and diarrhoea while away were associated with FRE carriage (p<0.05). 327 patients had travelled to a developing country of whom 53 carried FRE. A naive classifier based on this would mean screening 30% to detect 50% of FRE carriers. Using leave-one-out the best classifier meant 80% of FRE was detected if 87% of patients were screened.
Conclusions: Travel to a developing country was associated with an increased risk of carrying FRE. However no model could be constructed that would allow screening of a small enough high risk group that was sufficiently useful. Based on this all patients should be screened for FRE prior to a TRUS prostate biopsy.
Source Of Funding: Waikato Urology Research Limited
Waikato Hospital Hamilton NZ