Curtis Donskey’s group recently published a multicentre randomised trial in 16 US hospitals to evaluate the impact of an enhanced cleaning programme (including fluoruescent markers, environmental cultures, and feedback to cleaners) on the transmission of C. difficile. The intervention resulted in an increase in the removal of fluorescent markers, a reduction in environmental contamination with C. difficile, but no reduction in healthcare-associated CDI!

The 16 hospitals in Ohio were randomised 1:1 to the intervention (8 hospitals, but one dropped out so 7) and control (8 hospitals), stratified by the prevalence of CDI. Environmental cultures post discharge of high-touch surfaces in CDI and non-CDI rooms were performed once during baseline and once the intervention at each study site.

It’s worth nothing at the outset that hospitals were already using a sporicidal disinfectant for disinfecting CDI rooms (and 4/8 control hospitals, 2/7 intervention hospitals used bleach for terminal disinfection of all rooms). So, this study does not really tell us whether or not the contaminated environment is important in the transmission of C. difficile!

The intervention worked, in that the proportion of sites cleaned increased significantly (from 63% at baseline to 82% at the end of the intervention for post-discharge cleaning in CDI and non-CDI rooms), and the proportion of rooms contaminated with C. difficile decreased significantly (from a whopping 13% of rooms contaminated at baseline to 3% during the intervention). No significant changes in the proportion of rooms contaminated with C. difficile occurred. Despite these environmental findings, there was no significant reduction in hospital-onset healthcare-associated CDI, with the rate remaining static at round 6 cases per 10,000 patient day throughout the study period at the intervention and control hospitals (see figure).

Figure: Rate of CDI in control (top panel) and intervention (bottom panel) hospitals.

Perhaps the reason for this may simply be that the level of environmental reduction was not large enough to reduce the risk of transmission? Could this suggest that automated room disinfection systems – especially HPV – which go further than manual methods in reducing contamination from surfaces, may be required to achieve the level of environmental reduction necessary to impact upon environmental transmission of CDI? By the authors own admission, the culture methods used in this study were relative insensitive, so considerably more C. difficile spores may have been hanging around than were detected.

Perhaps the cleaners ‘got wise’ to the intervention and spot cleaned the fluorescent markers (which would also explain the environmental reductions since these same sites were sampled microbiological)?

Perhaps the focus on high-touch sites meant that other items in the rooms (mobile medical equipment, and floors) remained contaminated, negating the benefit of improved cleaning of high touch sites?

Perhaps hospital acquisition of C. difficile is actually a very rare event, and most apparent in-hospital acquisition is endogeneous?

Perhaps we need to screen and isolate carriers of C. difficle to successfully manage the contaminated environment?

Perhaps the fact that 6 control hospitals and 4 intervention hospitals already did fluorescent markers / ATP before the intervention started confounded the study?

Perhaps there was huge variability in other key IPC practices (such as antimicrobial stewardship and hand hygiene) that confounds the study?

Perhaps the rather limited modelling of the CDI rate (which did not account for any patient-level variables between hospitals) explains the clinical findings?

Or perhaps, as the authors conclude, an environmental intervention in isolation may not reduce CDI transmission, but can support a bundled approach reinforcing of other key IPC practices including antimicrobial stewardship and hand hygiene.

Finally, just to highlight what a great study this is – a huge undertaking for the research team, and just the sort of large randomised study that we need for evidence-based decision making in IPC.

ps make sure you check out Dan’s analysis of this study over on Controversies.