Only 18% of the hospitalized carriers had at least one surface of their environment contaminated, half of them being colonized with VRE. Considering only carriers of Enterobacteriaceae MDRO species, the environmental contamination rate was as low as 12.5%. This result is consistent with some recent studies, also performed in a non-outbreak context. (14,24,25). In contrast, 11 carriers of VRE among 26 (42%) contaminated at least one surrounding environmental site. Thus, the statistical analysis revealed that the carriage of VanA is a risk factor associated with environmental contamination. This high rate of environmental contamination around carriers of VRE is in line with many previously published studies. (7) Also, only one study found a correlation between the relative abundance of VRE in feces and the percentage of positive environmental samples found. (9) Therefore, we confirm the importance of the strict infection control policies’ application around VRE carriers. However, carriers of E. coli, between all other MDRO species and OXA48, among all MDRO’s mechanisms of resistance, seem to colonize significantly less the patients who contaminated their environment. The difference observed between VRE and Enterobacteriaceae MDRO species could explain the high rate of acquisition and occurrence of secondary cases around VRE carriers, widely described in literature. (9,26,27)
The most frequently contaminated environmental sites were bed sheet at the crotch level, pillow sheet and toilet seat. This finding was also previously reported. In fact, several studies found that the detection rate of different MDRO is reduced with increased distance from the carrier, with the bed surfaces being the most contaminated sites. (10,28,29)
Identifying patient factors associated with environmental dissemination within a hospital structure would make it possible to characterize those considered as high-level disseminators, allowing to better target environmental cleaning and minimize the risks of transmission. Risk factors could theoretically be related to the bacterial species itself and its ability to produce biofilms and resist, to factors associated with the host patient, such as a high degree of dependency or fecal incontinence, and/or elevated rectal abundance of MDRO. (11) Our study was able to identify only one factor clearly correlated with the risk of environmental dissemination, which is the carriage of VanA. However, three other factors were found protective against environmental contamination: having a urinary catheter, the carriage of E. coli between all other MDRO species and the carriage of OXA48. Interestingly, we found a high variability in the relative fecal abundance of all MDRO, but no correlation between the degree of gastrointestinal carriage and environmental contamination.
Curiously, when considering only MDRO Enterobacteriaceae species, we found no correlation between environmental contamination and any clinical characteristics of the carriers, particularly the load of gastrointestinal carriage, the Enterobacteriaceae species type and mechanism of resistance. Thus, none of the studied factors was associated with a significantly higher risk for environmental contamination, unlike Lerner and al., who found that high gastrointestinal concentration of CPE and fecal incontinence are risk factors for the environmental spread. (11) This contradictory result may be explained by the possibility that the dissemination of Enterobacteriaceae species in the environment is rather related to the bacterial species itself than its mechanism of resistance.
On another hand, not finding a statistically significant difference in environmental dissemination between carriers of E. coli MDRO and carriers of other Enterobacteriaceae MDRO species, being mainly K. pneumoniae species, stands in contradiction to different published studies demonstrating that the contamination is more frequent in the environment of Klebsiella carriers than in the environment of E coli carriers. In fact, those studies suggest that Klebsiella spp., known to form biofilms, which may be a way of surviving during long periods in the environment, have a higher persistence capacity in the environment that could account for the higher rate of cross transmission and the high potential to cause outbreaks in healthcare settings. (14,15,30,31) Our result could be explained by the possibility that the particular cleaning practices performed at the participating hospitals were adequate and had similar impact on K. pneumoniae and E. coli. However, unlike all previously published studies, our study took into consideration the patients’ characteristics and individual risk factors for environmental dissemination. In fact, patients’ characteristics profiles, including Charlson’s score of comorbidities and Kat’s score of dependence, were similar between carriers of E. coli and carriers of other MDRO Enterobacteriaceae species, being mainly K. pneumoniae. Since differences in risk factors’ profiles between patients with ESBL- E. coli and ESBL- K. pneumoniae, have been demonstrated by Freeman et al., patients’ individual risk factors could be confounding factors if not taking into consideration when studying environmental dissemination. (32)
Furthermore, we also haven’t found any statistically significant difference in environmental dissemination between carriers of ESBL-PE and CPE. Despite the fact that antibiotic use seemed to be more frequent in the group of ESBL-PE carriers, we would have imagined that it could lead to an increase in fecal carriage, resulting in a more important environmental dissemination, which wasn’t the case. Besides, according to the 2013 French national guidelines (33) and the international guidelines, (34) a strict isolation policy is applied on CPE carriers in healthcare facilities, including cohorting patients in a dedicated ward with dedicated healthcare workers and an extensive screening policy of contact patients. In case of identifying a non cohorted index patient, recommendations are to close the ward concerned and apply a screening policy to all contact patients. This strategy seems to be costly as is associated with bed closures and reduction of medical activity, for a relatively long duration of time. In addition to that, it could expose the isolated patients to a higher risk of complications as they are possibly receiving less optimal management for their medical condition, compared to non-isolated patients with the same medical condition. More recently, studies have suggested that neither single rooms, nor additional contact precautions, are necessary to control the spread of ESBL-PE. (35) Since we haven’t found any statistically significant difference in environmental contamination between carriers of CPE and ESBL-PE, and knowing that CPE involve the same bacterial species as ESBL-PE and the resistance genes are also plasmid-mediated, plus, in the absence of randomized studies demonstrating the mandatory nature of cohorting recommendations, we may wonder about the usefulness of these costly policies around CPE.
The major strengths of our work are that it was conducted in a non-outbreak context and in two different hospitals. Also, our study has proposed to assess the overall risk of environmental dissemination of MDRO, by evaluation the contamination of the environment, taking into consideration the individual risk factors together with the microbiological aspect of resistance. We acknowledge our study has some limitations. First, the small sample size, with the limited number of patients who have contaminated their surrounding environment. Second, we did not audit compliance with cleaning practices during the study. Third, although our swabbing technique was carefully standardized, and equated with a widely-used standard, it may not be the optimal means of detecting MDRO on surfaces.