The New England Journal of Medicine recently published an article evaluating the burden of CDI in the USA. The huge CDC-led initiative collected data from 10 geographically distinct regions, identifying more than 15,000 cases. Around two-thirds of cases were classified as healthcare-associated (although only 25% were hospital-onset). This means that, prima facie, a third of CDI cases were community-associated. I find this proportion difficult to believe: I strongly suspect that many of these cases would have had healthcare-associated risk factors if the team were able to look hard enough. For example, they used a fairly standard 12 week look-back period to evaluate previous hospitalisation, but how would the data look if they’d used 12 months? Also, it’s usually only possible to evaluate previous hospitalisation in a single healthcare system, but many patients commute between various healthcare systems. The authors acknowledge in the discussion that this designation of “community-acquired” may be inaccurate based on the finding from a previous study whether healthcare-associated risk factors were identified in most patients, but only be a detailed phone interview.

Scaling up from the figures from the 10 regions, national estimates were around 500,000 cases and 29,000 deaths due to CDI per annum in the US. This estimate is approximately double previous estimates for the national CDI burden in the USA, probably reflecting the adoption of molecular methods for the detection of CDI. This scaling up included an interesting statistical adjustment to see how prevalence varied depending on how many sites use sensitive molecular methods to detect CDI.

A sub-study included the culture of C. difficile from 1625 patients. More than 15% of stool specimens from patients diagnosed as CDI failed to grow C. difficile, probably illustrating the limitations of culture methods more than anything else. NAP1 (027) represented around half of cases, and was significantly more common in healthcare-associated CDI. I think it’s fair to say that the initial fears that NAP1 was a super-strain have been allayed by the fact that it’s now so common and there hasn’t been a surge in CDI mortality.

Finally, around 21% of healthcare-associated cases suffered at least one recurrence. Thus, there is a real need to the roll out of the uber successful faecal microbiota transplantation for recurrent CDI. In fact, there should be around 70,000 faecal microbiota transplantations each year in the US right now (500,000 x 0.66 x 0.21); I suspect there are far fewer.