Introduction Clostridium difficile infection (CDI) is the most common nosocomial infection in the United Kingdom. New treatments have shown reduction in CDI recurrence in clinical trials. Their high cost has been justified by savings on treatment of recurrence and reduction in financial penalties. However, in such a sick group of patients with multiple comorbidities recruitment rates to clinical trials are typically very low. A service development project was initiated to assess the current local burden of CDI and rates of recurrence.

Method For one year, between 30^thOctober 2013 and 2014, all samples tested positive for C. difficiletoxin production (Techlab Toxin AB Quikcheck) in the Microbiology department of Nottingham University Hospitals NHS Trust (NUH) were identified. Baseline age, leukocyte count, CRP, albumin and creatinine were recorded as well as use of proton pump inhibitors or other antibiotics if treatment was initiated by NUH. The main outcome measure was CDI recurrence within 6 months, with a particular focus on those treated by, and re-presenting to, NUH. At time of submission 6-month data is available for most cases with 3-month data available for all. In April 2014 the Trust adopted the use of Fidaxomicin for recurrent cases of CDI and for those receiving concomitant antibiotic treatment.

Results After exclusion of samples sent during the same episode (<20 days apart), the study identified 188 episodes of CDI in 164 individuals, mean age 71 ± 14. In 121 episodes this was the first presentation of CDI and there was a decision to treat medically (1 colectomy, 15 end-of-life). Of those 121, 16 patients had toxin (+) recurrence (13%). 3 patients had a 2^ndrecurrence, all of whom had been treated with Fidaxomicin on 1^strecurrence. 15 others received Fidaxomicin, 9 of whom had not responded to other treatment or recurred. There was no clear predictor of likelihood of recurrence. 32/121 patients died within 6 months of CDI (26%), of whom 23 died within 3 months (19%).

Conclusion CDI recurrence rates in medically treated patients were much lower than that reported for standard therapy in previous clinical trials (25%). Mortality in the medically treated group was much higher than in trials (6–7%) but similar to previous observational studies, reflecting under-representation of very sick patients in clinical trials. Treatment guidelines should recognise this limitation when applying clinical trial results to their patient population.

Disclosure of interest V. Wilkinson-Smith: None Declared, L. Koay: None Declared, T. Hills: None Declared, M. Diggle: None Declared, G. Major: None Declared, R. Spiller Grant/ Research Support from: Lesaffre, Ironwood, Norgine, Falk Pharma, Consultant for: Almirall, Astellas, Yuhan, Danone.