Systematic review: clostridium difficile and IBD

Abstract

Background: There is concern about an apparently rising incidence and worsening outcome of clostridium difficile infection (CDI) associated with inflammatory bowel disease (IBD). We have systematically reviewed, using PRISMA guidelines, the literature to evaluate the incidence, risk factors, endoscopic features, treatment and outcome of CDI complicating IBD. Method: Structured searches of Pubmed up to September 2010 for cross-sectional, cohort and case controlled studies. Results: Of 407 studies identified, 42 met the inclusion criteria: their heterogeneity precluded formal meta-analysis. There were no significant differences between the weighted mean incidence of CDI [weighted 95% CI] in adult patients with UC or Crohn's disease by either culture (UC 6.7% [5.6 11.4], Crohn's 8.8% [2.5 15.1]) or toxin assay (UC 3.1% [0.1 6.2], Crohn's 7.0% [1.8 12.2]). The weighted mean incidence of CDI, derived from two studies of children in relapse, was significantly higher than in adults (26.1% [16.1 36.1]). In remission, patients with IBD are more frequently colonised than otherwise healthy controls, the number of strains implicated suggesting multiple community sources of acquisition. Certainty about a temporal trend to increasing incidence of CDI in IBD, reportedly complicating 3.9% and 1.2% admissions for UC and Crohn's in 2004, increasing to 5.3% and 1.5%, respectively, in 2008, is compromised by possible detection bias and miscoding. Antibiotics are less of a pre-requisite for CDI in IBD than in the general population: where reported, only 43% [23/54] toxinpositive relapse samples were associated with antibiotic usage. Similarly, nosocomial acquisition (26% [5/19] IBD patients) is reported less frequently than in the general population. Further risk factors for acquisition, as in control populations, include increasing age, immunosuppressants and colonic disease. In 7 endoscopic studies, only 9% (13/150) of IBD patients with CDI had pseudomembranes. There are no controlled trials of CDI treatment IBD. Contradictory data, probably accounted for by differences in trial design, are reported for length of stay, colectomy and mortality rates. Conclusions: Patients with IBD seem to be at increased risk for CDI compared to controls and CDI in IBD has received increasing attention in the past decade, but whether its incidence is really increasing or its outcome worsening remains unproven. Endoscopy is unhelpful in diagnosing CDI in IBD. Therapeutic trials of CDI treatment are lacking and urgently needed.