Germs of thrones - spontaneous decolonization of Carbapenem-Resistant Enterobacteriaceae (CRE) and Vancomycin-Resistant Enterococci (VRE) in Western Europe: is this myth or reality?

In recent years, multidrug-resistant organisms (MDRO), especially Vancomycin-Resistant Enterococcus (VRE) and Carbapenem-Resistant Enterobacteriaceae (CRE) [1] are becoming more prevalent worldwide and represent a serious public health threat since they are becoming increasingly resistant to current treatment drugs. Moreover, CRE infections have been associated with an increased risk of mortality in Europe [2, 3].

In France, VRE and CRE belong to a group of Extensively Drug-Resistant (XDR) bacteria. They are spreading worldwide and oblige us to reinforce our in-hospital hygiene policy [4]. It has been suggested by the CDC [5] and national guidelines [6] that patients carrying such microorganisms must be ideally placed in isolation in a specific ward and cared for by a dedicated paramedical team, notably nurses. Nonetheless, it is well known that healthcare workers spent less time with patients in isolation which impacts negatively patient mental well-being [7].

Furthermore, no specific treatment is currently available to promote decolonization of these MDROs. Indeed, decolonization strategies have resulted in colonization relapse with the emergence of resistance to the intended therapies [8‐10]. More recently, fecal microbiota transplantation (FMT) has been proposed as a promising treatment option to clear XDR carriage in mice [11] and in humans [12‐14], including immunocompromised individuals [15]. To date, there are no randomized clinical trial data to support its efficacy, but previous cited findings suggest that FMT might be more effective against VRE than CRE colonization. Yet, as FMT in such indication is still under evaluation it cannot be performed in common clinical practice, thereby leaving physicians without immediate solution.

A recent systematic review by Bar-Yoseph et al. showed that in the natural history of colonization among healthcare residents, a significant proportion of CRE carriers (34.6%) remained colonized for up to 1 year [16]. While this meta-analysis is of interest, it is mainly composed of 4 cohorts which benefited simultaneously from active decolonization therapy and thus did not focus on spontaneous decolonization.

As data are limited in the literature, we decided to analyze the natural history of spontaneous decolonization in XDR colonized patients and identify whether XDR carriage was associated with a larger care burden with a difference from the standard patient care because of the cohorting strategy recommended in acute healthcare facilities [5].

In our present cohort, spontaneous clearance of XDR bacteria is not so rare and occurred in 48.2% of cases with a median time of 49 days and one-third of them were decolonized within a month. Those findings are concordant with the limited data available in the literature reporting natural clearance of MDRO and especially CRE [17‐22] and VRE [23, 24] carriers during hospital settings. All of them are retrospective data, mainly concerning KPC cases occurring in Israel, except one prospective sub-study issued in China by Cheng et al. [22] which focused on NDM-1 isolates and reported a surprisingly high rate of 72.2% (n = 57) of decolonization within a short period (median delay of 30 days). Furthermore, sample size ranges from 66 to 125 patients and some studies present mixed population by adding patients in rehabilitation centers and long-term care facilities.

To our knowledge, this is the first study that describes natural decolonization of XDR (CRE and VRE together) carriers in an adult medical ward in Western Europe (France). Actually, in the literature only one study was conducted in Germany but concerned an outbreak of Carbapenem-Resistant K. pneumoniae related to a patient transferred from Greece [17] where the prevalence is deemed to be high (around 60% in 2013) [25]. Even if the prevalence of CRE is still rare in France in 2016 (< 1%) [25], considering the boarders we share with the Mediterranean countries, we should be wary of possible outbreaks.

Also, our cohort illustrates 2 interesting data. Firstly, decolonization does not seem to result from the type of microorganism (VRE versus CRE in Table 3). This is consistent with the review of the literature, showing a similar median of decolonization (around 6 months) between VRE [24] and CRE carriers [16]. Secondly, there is a trend of persistent bacterial carriage with XDR bacteria under antibiotic exposure (p = 0.06), in the limit of the sample size. Therefore, we could not detail the different antimicrobials and prescriptions duration impact on decolonization. Nevertheless, we can assume that ongoing antimicrobial therapy may slow down the clearance of XDR bacteria over time, considering antimicrobial agents impact the gut microbiota [26]. This hypothesis is concordant with Schechner et al. who argued that prior fluoroquinolones use was predictive of persistent CRE rectal carriage [21] and with Cheng et al. who reported an impact of the use of proton pump inhibitors in addition to antimicrobial therapy [22].

We also observed 2 patients retested positive for CRE after 2 negative swabs during follow-up, but both received antimicrobial agents during the time interval. Such observation reinforces our belief that there is a close link between antibiotic exposure and selective pressure on gut microbiota.

It is important to note that decolonized patients had longer LOS than colonized patients (p = 0.02). This may be explained by the fact that decolonized patients had also longer follow-up (p < 0.0001). Also, while patients are colonized, physicians are under pressure to expedite patient discharge to avoid in-hospital spread of XDR bacteria. Moreover, their follow-ups are difficult to undertake, because most of the patients returned back home (72.8%) mainly because of a refusal of other healthcare facilities to set up necessary precautions. Therefore, a majority of patients did not undergo screening after 1 year (85.6%).

Moreover, there is a lack of a consensual definition of clearance, including VRE, partly due to heterogeneity in the studied populations [24]. In the present work, we considered at least 2 negative rectal swabs (> 1 week) as a baseline for the decolonization process while Lubbert et al. proposed a series of at least 4 consecutive negative rectal swabs (> 1 week) [17], which is one major point that limits the interest of all the above findings in real-life conditions. In 2017 in the 39 hospitals of AP-HP, the central committee for hygiene policy (CLIN) proposed a series of at least 3 negative swabs with one-year hindsight before considering the patient free from XDR bacteria. Yet in Garches, we reduce precautions from quarantine to contact precautions after 4 months of negative swabs, considering that patient might still excrete XDR bacteria as soon as he is exposed to antibiotics. Such long period of time can be sometimes really hard to face the patients, especially when they are not infected.

Therefore, we believe our findings might be of interest to guide hygiene policies accurately about reducing extra precautions on readmission to the healthcare system to lower a significant burden on healthcare provision. Besides, one solution to enhance bacterial clearance might be the FMT proposed by our team as a “fecal weapon” which showed promising results [14].

Finally, this study has several limitations to be considered. First, we experienced a high rate of people lost to follow-up after 1 year. This can be explained by the fact that once a patient has been discharged, it is very difficult to obtain a monthly rectal swab without evidence-based medicine. Second, we did not perform systematically PCR testing on follow-up swabs, considering that PCR was only legitimate for a rapid initial diagnosis of carriage and not indicated in routine surveillance considering its low positive predictive value of 16.6% [27] and would result in higher costs. Thirdly, whereas multiple hospitalizations (198 stays for 125 patients) were recorded in our analysis, we cannot warrant that some patients were hospitalized in other acute care facilities between the initial culture and the final follow-up.

In a context of multidrug resistance development and a lack of new antimicrobial agents, a stringent surveillance of XDR bacteria is needed. Our study shows that a spontaneous decolonization occurred within the first 30 days in 16.4% of cases and up to 48.2% after day-90. This major point confirms that spontaneous decolonization is not so rare and does not seem to be static over time which is very encouraging. We observed a larger care burden in almost one-tenth of XDR carriers, because of their quarantine condition. This must be considered when physician decides to screen for a XDR bacteria. Finally, we can regret that a majority of patients did not undergo screening after 1 year, emphasizing the need for more monitoring and prospective studies.