SUMMARY Twenty three cases of pseudomembranous colitis (PMC) occurred in three hospitals in 10 months. Retrospective analysis shows that they represented a single epidemic with a readily traceable chain of person-to-person contact within and between hospitals. Most patients had severe pre-existing illness and all had broad spectrum antibiotics, including either ampicillin/ amoxycillin, a broad spectrum cephalosporin (particularly cefotaxime), or both. All patients had severe diarrhoea and all responded to vancomycin, but relapse occurred in five. Ten patients eventually died, principally because of underlying disease rather than from PMC. Failure to find fibrin thrombi in blood vessels in biopsies and the scanty distribution of non-invasive bacteria supports the concept of mucosal damage by bacterial toxin, rather than by direct infection, or ischaemia. Although environmental colonisation cannot be excluded, the observed pattern of spread suggests a major role for direct person-to-person crossinfection in the spread of disease in this outbreak.Pseudomembranous colitis (PMC) is an inflammatory condition of the colon, usually occurring after broad spectrum antibiotic therapy, and associated with colonisation by toxigenic Clostridiium difficile.We report an epidemic of PMC in three hospitals of our hospital group, in which there was temporal and geographical clustering of cases suggestive of crossinfection. C cdifficile is generally considered an opportunistic pathogen which may cause PMC in patients who have received broad spectrum antibiotics. This outbreak, taken in conjunction with other recent reports of case clustering,"' emphasises the role of cross infection in the pathogenesis of PMC. Reccied for puhblication IS April 1987. had also been identified in hospital 2. A retrospective study was started. Clinicians in all seven hospitals of the group were questioned (in particular all endoscopists) about cases of PMC occurring in the previous year. Histopathological surgical and autopsy records for the previous year were reviewed for cases diagnosed as PMC and all microbiological isolates of C difficile were recorded. Hospital charts of all patients identified by these procedures were reviewed and those not having a clinical illness consistent with PMC were excluded. In addition to 15 cases initially known to the clinicians at hospitals I and 2 a further eight cases were identified by this screening procedure. Twenty patients in whose faeces C difficile was cultured showed no clinical evidence of PMC and these, regarded as asymptomatic colonisation, were eliminated from the study. We feel it unlikely that any diagnosed cases of PMC within our hospital group were not included in the study.After these investigations it was found that the cases of PMC were limited to three hospitals, all of which are members of a larger hospital group, served by a central laboratory. Each affected hospital has 1467 on 11 May 2018 by guest. Protected by copyright.
