Introduction Severe Clostridium difficile associated diarrhoea (CDAD) is an important nosocomial infection, often resulting in severe morbidity or death. The rates of CDAD have increased significantly in the last 2 decades, but predictors of outcome are poorly understood. Methods A retrospective cohort study was performed in patients with a diagnosis of CDAD hospitalised at Leeds Teaching Hospitals NHS Trust (LTHT) between January 2011 and December 2011. The data on these cases was collected from electronic patient records and medical notes. Data collected included general demographics, underlying medical conditions, Horn Index, Charlson co-morbidity score, clinical and laboratory data, and the medical treatment given. Death due to any cause either during that hospital stay or within 30 days of discharge from hospital was the primary outcome. Severe CDAD was defined according to the UK Health Protection Agency (HPA) guidelines as WCC > 15 X 109/L, or an acute rising serum creatinine (i.e. > 50% increase above baseline), or a temperature of > 38.5°C, or evidence of severe colitis (abdominal or radiological signs). Logistic Regression analysis was used to identify parameters associated with mortality. SPSS version 17 (IBM Corp, NY) was used to perform the statistical analysis. Results There were 247 patients with a diagnosis of CDAD made in 2011 at LTHT of which 16 were wrongly coded, 5 were treated in the community, 12 had insufficient information in the notes and in 68 patients the medical notes could not be traced. A total of 170 episodes in 146 patients were finally analysed. There were 36 deaths in this group. Patients who were dead were older (mean age 78±12.9 vs 76.6±17.6). Independent predictors of mortality on multivariate analysis included age (OR 1.051, 95% CI 1.009–1.095), Charlson co-morbidity score of ≥3 (OR 3.036, 95% CI 1.209–7.622), Horn Index (Major or Extreme) (OR 4.725, 95% CI 1.818–12.283), Severe CDAD (OR 3.454, 95% CI 1.222–9.760) and in-appropriate treatment of severe CDAD with metronidazole as first line therapy (OR 4.642, 95% CI 1.213–19.193). Factors not found significant included gender, prior use of antibiotics, PPI use, opioid use, prior episodes of CDAD and treatment with vancomycin. Conclusion Predictors of all-cause mortality in patients with CDAD include older age, Charlson score≥3, Horn index ≥3, severe CDAD as defined by the UK HPA and in-appropriate use of metronidazole in severe CDAD. Patients with severe CDAD should not be treated with Metronidazole as first line therapy. Further prospective validation of these results is needed in a multicenter setting. Disclosure of Interest None Declared.