Comparison of Control of Clostridium difficile Infection in Six English Hospitals Using Whole-Genome Sequencing.
Eyre DW., Fawley WN., Rajgopal A., Settle C., Mortimer K., Goldenberg SD., Dawson S., Crook DW., Peto TEA., Walker AS., Wilcox MH.
Background: Variation in Clostridium difficile infection (CDI) rates between healthcare institutions suggests overall incidence could be reduced if the lowest rates could be achieved more widely. Methods: We used whole-genome sequencing (WGS) of consecutive C. difficile isolates from 6 English hospitals over 1 year (2013-14) to compare infection control performance. Fecal samples with a positive initial screen for C. difficile were sequenced. Within each hospital, we estimated the proportion of cases plausibly acquired from previous cases. Results: Overall, 851/971 (87.6%) sequenced samples contained toxin genes, and 451 (46.4%) were fecal-toxin-positive. Of 652 potentially toxigenic isolates >90-days after the study started, 128 (20%, 95% confidence interval [CI] 17-23%) were genetically linked (within ≤2 single nucleotide polymorphisms) to a prior patient's isolate from the previous 90 days. Hospital 2 had the fewest linked isolates, 7/105 (7%, 3-13%), hospital 1, 9/70 (13%, 6-23%), and hospitals 3-6 had similar proportions of linked isolates (22-26%) (P ≤ .002 comparing hospital-2 vs 3-6). Results were similar adjusting for locally circulating ribotypes. Adjusting for hospital, ribotype-027 had the highest proportion of linked isolates (57%, 95% CI 29-81%). Fecal-toxin-positive and toxin-negative patients were similarly likely to be a potential transmission donor, OR = 1.01 (0.68-1.49). There was no association between the estimated proportion of linked cases and testing rates. Conclusions: WGS can be used as a novel surveillance tool to identify varying rates of C. difficile transmission between institutions and therefore to allow targeted efforts to reduce CDI incidence.