Failure of standard precautions

Standard precautions are the basis of all infection control measures. Their importance has been reinforced by an interventional study looking at different interventions to reduce colonization and transmission of MDR bacteria in intensive care units (ICU).^{Reference Derde, Cooper and Goossens23} Improved hand hygiene plus unit-wide chlorhexidine body washing reduced the acquisition of antimicrobial-resistant bacteria, particularly MRSA, but additional screening and isolation of carriers did not further reduce acquisition rates. However, when analyzing the subgroup of multidrug-resistant Enterobacteriaceae, none of the interventions resulted in a reduction of acquisition, suggesting that in addition to standard precaution, even contact precautions may fail to contain multidrug-resistant Enterobacteriaceae and further considerations as detailed in the following may need to be considered.

The transmissibility of antibiotic-resistant Enterobacteriaceae in an ICU setting has been estimated to be nearly 4 times higher for non-EC Enterobacteriaceae than for EC.^{Reference Gurieva, Dautzenberg, Gniadkowski, Derde, Bonten and Bootsma24} One reason that MDR-KP is more transmissible than other Enterobacteriaceae may be its ability to survive in a hospital environment. This survival potential was shown in a pediatric population where 470 environmental samples were taken. Among them, 19 yielded ESBL-producing bacteria.^{Reference Guet-Revillet, Le Monnier and Breton25} Klebsiella pneumoniae was the most frequently isolated species, and ESBL-KP carriage or infection was the only risk factor associated with surface contamination in this trial. These findings have been further supported by a study conducted in France, where contamination of the environment was significantly higher in patients colonized or infected with ESBL-KP than other ESBL-producing Enterobacteriaceae (ESBL-PE).^{Reference Gbaguidi-Haore, Talon, Hocquet and Bertrand26} Such data support the proposition that infection control measures need to be tailored to the different members of the family of Enterobacteriaceae.

Late identification of carriers

In the German outbreak,^{Reference Kaiser, Finstermeier and Hantzsch21} the index patient was identified 10 days after hospital admission. Late identification of carriers has also been understood as a key issue, triggering an outbreak of KPC-KP occurring in Finland in 2013 involving 9 patients.^{Reference Kanerva, Skogberg and Ryynanen27} The index case was not identified for almost 4 months because this patient did not feature any risk factors known for MDR-KP such as travel abroad.

Early identification of carriers is key for timely implementation of contact precaution measures to prevent ongoing transmission, but how can we improve early detection of carriers? Although the implementation of screening protocols seems to be the obvious answer, many questions around screening remain unanswered at this point, including which patients need to be screened, how screening is to be performed, and at which time point screening is most specific.

Who to screen. The European Center for Disease Control (ECDC) published guidelines in 2017^{Reference Magiorakos, Burns and Rodriguez Bano15} providing recommendations on who to screen by asking several key questions at hospital admission. These questions include (1) history of an overnight stay in a healthcare setting within the last 12 months, (2) whether the patient been dialysis-dependent or received cancer therapy within the last 12 months, (3) known history of MDRE, and (4) whether the patient epidemiologically linked to a carrier of MDRE. If any of these answers is yes, the patient is to be placed under preemptive contact precautions and screened. Although these recommendations are reasonable, few institutions have the capacity for enough patient isolations when most admitted patients are affected. More restrictive screening and preemptive isolation recommendations may result in lower sensitivity with improved feasibility, resulting in a better cost–benefit ratio.

In the United States, active screening of patients with known risk factors for carriage of MDR organisms is suggested by the Centers for Disease Control and Prevention (CDC), yet no specific guidelines similar to those published by the ECDC exist.¹⁷ According to the 2006 Healthcare Infection Control Practices Advisory Committee (HICPAC) guidelines for management of multidrug-resistant organisms in healthcare settings, it is advisable for acute-care facilities in nonendemic regions (1) to screen microbiological records of the 6–12 preceding months for carbapenem-resistant Enterobacteriaceae (CRE), (2) to perform point prevalence culture surveys in high-risk areas in case unrecognized CRE cases occurred to identify subsequent cases of CRE, and (3) to perform active surveillance cultures of patients with epidemiologic links CRE cases.^28,29

Where to screen. To reduce false-negative screening results, the most frequently colonized body sites should be screened, yet data investigating which sites to screen are limited. When systematically screening colonization sites for ESBL-PE, urine was the site mostly colonized (83% of all cases), followed by rectum (69%), groin (35%), and throat (13%). Among this population, 24% of all carriers were only detected as such by screening the urine; however, most ESBL-PE (ie, 86%) were E. coli, thus limiting generalizability to other Enterobacteriaceae.^{Reference Tschudin-Sutter, Frei, Dangel, Stranden and Widmer30} In a subsequent study, the rectum was not colonized in 14% of all ESBL-PE carriers without reporting differences between different Enterobacteriaceae.^{Reference van Prehn, Kaiser, van der Werff, van Mansfeld and Vandenbroucke-Grauls31} Among patients in Thailand, a high prevalence of ESBL-PE colonization in rectal specimens followed by groin, throat, and nasal cavity was shown, yet urinary samples were not investigated.^{Reference Rattanaumpawan, Choorat, Takonkitsakul, Tangkoskul, Seenama and Thamlikitkul32} Considering that most guidelines recommend screening by performance of rectal swabs only, a notable number of colonized patients might remain undetected when urinary sample screening is not additionally performed. Further studies are needed to confirm sites needing to be considered by screening protocols to increase sensitivity.

When to screen. A recent study found that patients screened on admission had a 6% probability of being positive for KP, whereas patients screened at least 3 days after admission had a probability of 19%.^{Reference Gorrie, Mirceta and Wick33} Thus, patients may also need to be screened after admission and repeatedly during a hospital stay and not just at hospital admission, as recommend by most guidelines. Although detection of acquisition in the hospital might play a role, selection pressure under antibiotic treatment seems more likely.

How to screen. Minimizing the interval between patient admission and initiation of contact precaution is a crucial factor in infection control. However, turnaround time of screening results and time to patient isolation remain major challenges.

A recent study by Ducomble et al^{Reference Ducomble, Faucheux and Helbig1} investigated whether polymerase chain reaction (PCR) could be performed directly from rectal swabs without waiting for cultures, which could decrease time to contact isolation precautions from 5 to 1.5 days, turnaround time from 8 days to 1.5 days, and confirmed patient days in isolation of patients with KPC-KP increased from 72% to 90%.

In a study by Oueslati et al,^{Reference Oueslati, Girlich, Dortet and Naas34} a PCR-based test performed directly from bacterial colonies detected the carbapenemases mostly found in MDR-KP with high sensitivity and specificity within 2 hours, improving turnaround time, which might lead to faster identification of antibiotic resistant strains and thus not only reduce time to isolation but also reduce inadequate antimicrobial treatment.

Silent reservoirs

The establishment of silent reservoirs is another key issue to consider regarding transmission of MDR-KP. Contamination of a positioning cushion, possibly endoscopes and mobile hemodialysis machines, were considered crucial in the continuation of the German outbreak.^{Reference Kaiser, Finstermeier and Hantzsch21}

A French study reported an outbreak entertained by similar reservoirs.^{Reference Pantel, Richaud-Morel, Cazaban, Bouziges, Sotto and Lavigne35} MDR-KP survived on a mattress but also in sinks leading to ongoing transmission. The transmission of resistance mechanisms not only occurs within the species itself but also in transmission of resistance genes to other pathogens. This phenomenon was revealed by Conlan et al,^{Reference Conlan, Thomas and Deming36} who found evidence of horizontal transfer of carbapenemase-encoding plasmids between KP, Enterobacter cloacae, and Citrobacter freundii in the hospital environment.

Endoscopes are an important reservoir to consider in an outbreak investigation. An OXA-204 carbapenemase Enterobacteriaceae outbreak was related to endoscopes in France.^{Reference Potron, Bernabeu and Cuzon37} The outbreak involved 2 different species, KP and EC. Among all 22 patients affected by the outbreak, among the clusters of KP isolates, which were transmitted by endoscopes, patients actually went on to transmit their KP to others by patient-to-patient transmission, whereas the EC isolates were no longer transmitted. These results support the findings on superior transmissibility of KP compared to other Enterobacteriaceae.

Interventions targeting wastewater hospital plumbing fixtures by installing hopper covers and sink tap devices in adult ICUs, demonstrated a decrease in patients acquiring KPC, pointing to the hospital wastewater as an important source for transmission.^{Reference Mathers, Cox and Kitchel38}

Removal of sinks and introduction of water-free patient care was shown to decrease the rates of gram-negative colonization in an ICU after discovery of an outbreak of ESBL-producing E. cloacae.^{Reference Hopman, Tostmann and Wertheim39} Colonization rates dropped significantly from the preintervention period to the postintervention period, most clearly in patients with increased lengths of stay in the ICU. These findings indicate that a way to enhance control of MDR-KP might be by introduction of water-free patient care.

No-touch disinfection systems may be an effective tool to prevent establishment of silent reservoirs.^{Reference Ali, Muzslay, Bruce, Jeanes, Moore and Wilson40} Overall, 95% of 214 sites were contaminated with bacteria after manual terminal disinfection, with high numbers of bacteria being present on the floor, bed control panels, and the nurse call buttons. Enhanced disinfection using hydrogen peroxide vapor aerial decontamination systems reduced surface contamination to low levels. Yet introduction of these systems leads to prolongation of turnover time and higher costs.

Transfer within and between institutions

A complex exposure network summarizing pathways of a moderate amount of patients maintaining an outbreak of KPC Enterobacteriaceae through numerous transfers between nursing homes, acute-care institutions, and long-term acute-care hospitals revealed the emergence and rapid regional spread of MDR-KP.^{Reference Won, Munoz-Price and Lolans41}

In another study, patients hospitalized in geriatric care wards were systematically screened for carriage of KP.^{Reference Gorrie, Mirceta and Wick42} Unique different lineages of KP in the majority of patients were detected, and these lineages were susceptible to most antibiotic agents. However, they also found clusters of MDR-KP strains with minimal nucleotide divergences. These clusters were attributable to ESBL producers carrying bla_{CTX-M-15 ,} which are clones associated with hospital-acquired ESBL infections worldwide. Linking these data with the epidemiologic data suggested that these strains were actually transmitted within the referral hospital, indicating that screening of transfer patients might be a solution for MDR-KP infection control, yet viability of these procedure seems to go beyond the scope of hospital capacity looking at the immense frequency of interfacility referrals, which again underscores the need for amended screening algorithms.

The number of patient transfers between facilities and the number of genetic links has also been modeled in a study investigating an outbreak of MDR-KP. Phylogeny indicated that a strain of MDR-KP was brought into each facility multiple times by various intrahospital transmission routes, suggesting that patient sharing perpetuated the regional outbreak.^{Reference Snitkin, Won and Pirani43} Consequently, the density of patient transfers alone is predictive of interfacility transmission. Thus, implementation of pathogen surveillance by whole-genome sequencing might help to identify facilities within a patient-sharing network and lead to targeted investigation and infection control intervention.

Despite this knowledge, only 70% of all laboratories in the United Kingdom flag positive results of MDR-KP on patient records, and only 30% of laboratories feature a system in place for alerting neighboring trusts.^{Reference Trepanier, Mallard and Meunier44} Protocols for transfer from hospital including MDR-KP notification are only available for 75% of laboratories.

Clear communication about the MDR-KP status of patients at transfer between facilities is crucial so that each institution is able to conduct infection control measures at hospital admission directly. Rosenman et al^{Reference Rosenman, Szucs and Finnell5} provided data of an automated system to notify healthcare workers in case a patient with a known history of an MDR-organism was admitted to an emergency room or an inpatient setting. The automated system recorded substantial crossover of affected patients between institutions within a short time and provided information to help reduce time to isolation. Notably, these findings rest on the premise that we have a clear definition of MDR organisms.^{Reference Rosenman, Szucs and Finnell5}