Ever found yourself wondering how often colonization precedes infection with MDR-GNB and VRE? A new systematic review and meta-regression in Lancet ID gives us a pretty solid answer: about 14% of MDR-GNB and 8% for VRE. This information is helpful for us to qualify the significance of acquiring these organisms in healthcare settings.

There was always a rough rule of thumb for a 1:10 ratio between MRSA colonization and infection. But I wouldn’t have been able to guess what this would be for MDR-GNB and VRE. So, found the recent Lancet ID review and meta-analysis extremely helpful! A total of 40 studies were included in the meta-regression (more info on what a meta-regression is here). 32 of the studies included MDR-GNB and 16 included VRE. The key findings were that 14% of patients with MDR-GNB developed infection during the followup period, which was significantly more than VRE at 8%. Within the MDR-GNB group, CRE had the highest proportion at 19% compared with 9% for Klebsiella and 8% for E. coli and for Gram-negative bacteria with ESBLs. The corresponding figure for bloodstream infection were 7% for MDR-GNB overall (10% for CRE, 4% for ESBL and E. coli, and 2% for Klebsiella) and 10% for VRE. CP-CRE showed an even higher risk than for CRE.

The cumulative incidence of infection did not vary by specific patient populations for MDR-GNB (ICU, cancer, transplant) but did for VRE, with patients in ICU, cancer, or transplant areas having cumulative infection incidence rate than the general hospitalised patient.

It seems difficult to explain the 10 percentage point difference between CRE and other multidrug-resistant Enterobacterales. Whilst the authors attribute this to increased virulence of microbes that acquire carbapenem-resistance, this seems unlikely and would go against another “rule of thumb” that microbes generally make a trade-off between acquiring resistance and the ability to cause infection. I wonder whether it has more to do with study design and/or differences in surveillance methods for CRE vs. non-CRE?

Some limitations to consider:

• These studies would have included a mixture of “prevalent” colonization, and “hospital-acquired” colonization, and it would be interesting to see whether the infection : colonization ratio was different for these.
• The studies included quite a wide range of followup periods, which would influence the transition from colonization to infection.
• There’s some interesting geography in where the studies were performed: most VRE studies were performed in the USA, and most CPE studies in Europe. It isn’t clear how / whether this would influence the findings.
• Whilst all of the studies included gastro-intestinal colonization, some also included colonization at other body sites, which may skew findings.

This study is really helpful but actually rather alarming – to think that something like 1 in 6 or 7 patients with MDR-GNB colonization will develop an infection in the next 30 days or so. And even worse – to think that this will be 1 in 5 for CRE, and 1 in 10 of patients colonized with CRE will develop a CRE BSI in the next 30 days! This certainly reinforces the need to prevent the spread of MDR-GNB, and especially CP-CRE, in our hospitals.