This study describes the retrospective detection of a DAI caused by a normal susceptible microorganism and two duodenoscopes contaminated with the same microorganism. Due to the retrospective nature of this study, the exact route of transmission between the two duodenoscopes could not be clarified. This ERCP center uses a more extensive surveillance system than the twice a year surveillance prescribed by national guidelines to detect contaminated duodenoscopes [6]. Duodenoscope surveillance cultures were collected monthly and in the case of duodenoscope 1, even three times a month during the contamination period. However, even despite this more intense surveillance, contamination of the described duodenoscopes was missed. The study cultures revealed at least five moments duodenoscope 1 was contaminated with E. cloacae complex that were missed by surveillance. This DAI would only have been avoided if this duodenoscope was cultured on a daily basis while being quarantined awaiting culture results.
In recent years, a higher level of awareness regarding contaminated duodenoscopes and subsequent serious infections was noticed as demonstrated by literature reports of multiple incidents [1]. However, DAIs are underreported as this needs elaborate and specific microbiological culture and surveillance actions. For improvement three prerequisites are needed. First, cultured microorganisms of endoscopes and patients need to be stored. Second, linking patient cultures to contaminated duodenoscopes should be easier, for instance with alerts in the electronic patient files in case a used duodenoscope is found to be contaminated. Third, if a match is suspected, molecular typing should be performed to prove indistinguishability of the microorganisms found in the patient and duodenoscope cultures. In this case, we started with molecular typing using AFLP, which is a relatively quick and affordable typing method. Only the isolates that were indistinguishable through AFLP were subjected to WGS. WGS is a more laborious and expensive method, but provides a more in-depth analysis of the genetics of the different microorganisms [7]. In most DAI outbreak investigations, molecular typing is performed using PCR techniques [8,9,10]. Only in a few outbreaks, WGS was used to distinguish the microorganisms found on the duodenoscopes and in the affected patients [11,12,13]. WGS is superior to other types of molecular typing such as AFLP or pulsed-field gel electrophoresis (PFGE) in strain typing [7, 14].
In this report, AFLP and classical MLST results showed the microorganism in the patients’ blood culture to be indistinguishable from isolate 4 grown from duodenoscope 1. They were found to be of a new MLST type, making transmission of the same type very plausible. However, WGS of the two microorganisms revealed 93 allelic differences in the 3757 identified alleles. There is no specific cut-off level to determine if both microorganisms could still be related despite these allelic differences [15]. Thresholds of maximum differences in WGS typing are species specific and are not clearly defined. In clinical practice, cut-off levels are usually set at ~ < 1% of the analyzed genes. However, predetermined cut-off values need to be used cautiously as they do not take into account situation specific characteristics. This includes the long interval between the acquisition of the samples from the duodenoscope and the patient, the method of storage of the distinct isolates and the disinfection of the duodenoscope prior to sampling. These characteristics might have influenced the WGS results of the specific microorganism. Also, although 93 different alleles sound like a large difference, it is not much compared to truly unrelated isolates. Upon analysis of other E. cloacae complex isolates from various origin, isolates that were considered related or unrelated showed either < 5 or > 2000 allelic differences, respectively (results not shown). The 93 allelic differences, together with the observation that these isolates represented yet unknown MLST types indicate a certain degree of relatedness and that these isolates might link to a common source. Therefore, in our case, WGS did not provide conclusive proof of a transmission, but all results taken together with the epidemiological information makes it plausible that there is a link between the contaminated duodenoscope and the patient infection.
Several factors explain why this DAI was not considered and recognized as such at the time of infection. First, the long time period of 106 days between the ERCP procedure and infection likely contributed to the fact that this infection was not recognized as a DAI in clinical practice. Second, as this infection was caused by a susceptible common gut microorganism, which is not a rare cause of cholangitis [16], it is understandable that the treating physicians did not consider the duodenoscope to be the source. Third, due to the patient’s underlying malignancy, he was treated on an oncology instead of a gastroenterology ward. Awareness regarding the possible occurrence of DAIs is likely to be lower or even absent among non-gastroenterological personnel compared to health care workers involved in ERCPs on a daily basis. Fourth, the microorganisms of interest were found in study cultures for which clinicians and infection prevention practitioners were blinded per study protocol. Concurrent surveillance cultures did not detect E. cloacae complex or other gut flora. This shows that monthly surveillance cultures are inadequate to detect DAIs.
No patients were found to be treated with both duodenoscopes in the period between the first confirmed contamination of duodenoscope 1 and 2, making it unlikely that one patient was responsible for the contamination of both duodenoscopes. Therefore, a contamination source in the reprocessing surroundings could explain how these two separate duodenoscopes became contaminated with the same microorganism. However, since our investigation took place long after the transmissions had occurred, no environmental samples were acquired to test this hypothesis. Cultures of the final rinse water of the automated endoscope reprocessors (AERs) collected in July and October were negative for growth of any microorganisms. No other duodenoscopes were contaminated with this specific strain. Both duodenoscopes had recently been investigated by the manufacturer. Duodenoscope 1 was sent to the manufacturer on August 9th to check an issue with the image quality, duodenoscope 2 was sent to the manufacturer on July 14th for periodic maintenance. Hypothetical explanations for a transmission between the two duodenoscopes could be that any of the cleaning materials (i.e. brushes or syringes) were used on both duodenoscopes instead of being disposed after every use or that the reprocessing staff served as a transmission vector due to incorrect use of protective equipment or incidental breaches in the reprocessing protocol. However, since the transmission took place during a study which was focused on reprocessing efficacy, it seems more likely that the reprocessing staff adhered very strictly to the cleaning protocols in this period.
Duodenoscope 1 contained two different strains of E. cloacae complex prior to the ERCP of the patient. The strain belonging to ECC cluster 1 was found in three samples of the working channel collected over a period of 2 months. The strain belonging to ECC cluster 2 was identified once and only found in a sample from the forceps elevator. Our hypothesis is that the strain from ECC cluster 1 was able to develop a biofilm in the channel, explaining its persistence over 2 months, whereas the presence of the strain from ECC cluster 2 was more likely to be an incidental event.
In 2012, our hospital was affected by a duodenoscope-associated outbreak which caused infections with a VIM-2 producing Pseudomonas aeruginosa in eight patients [17]. Ever since, we have been actively involved in research regarding duodenoscope contamination and DAIs. The current DAI experience has led to a few local changes in our hospital, such as the use of different brushes, borescope inspections and an intensified surveillance protocol. In other outbreak reports, the outbreaks were ended by more strict adherence to reprocessing and drying protocols [10], by removal of the duodenoscopes [18, 19] or by introducing ethylene oxide sterilization [19,20,21]. In our case, the contamination of duodenoscope 1 disappeared spontaneously after 3 months and the duodenoscope was already removed from service at the time of this retrospective investigation. However, since duodenoscope contamination remains a worldwide problem with approximately 15% of all patient-ready duodenoscopes being contaminated [22], structural solutions are needed to ultimately banish this problem. Ever since the first outbreak reports, a lot has changed in the reprocessing methods, with the introduction of AERs and drying cabinets. Also, several national and international guidelines have been developed over the years and duodenoscope manufacturers have altered the duodenoscope designs to improve cleaning accessibility [23]. The FDA has been actively involved in the topic since 2013 and has issued numerous recommendations and warnings on the use and reprocessing methods of duodenoscopes and has ordered the three main duodenoscope manufacturers to conduct postmarket surveillance studies [24]. Despite all this effort, a zero contamination rate has not yet been reached for reusable duodenoscopes. The only realistic option to completely prevent DAIs is switching to the recently introduced disposable duodenoscopes for single use. However, the substantial costs associated with that switch make it unlikely that reusable duodenoscopes will soon be replaced by disposable ones [25].
Although treatment of endogenous and exogenous post-ERCP infections are not necessarily different from one another, it is important to make this distinction. DAIs are preventable, but rare iatrogenic complications, whereas endogenous infections are in general inherent to the ERCP procedure. Since patients undergoing ERCP procedures are often vulnerable and immunocompromised, such as our patient described here, post-ERCP infection is a serious and potentially fatal complication [19]. Furthermore, one contaminated duodenoscope can remain contaminated over a long period of time and infect multiple patients [9, 10, 20, 26], as is demonstrated by outbreak reports in which 12–41% of the exposed patients became colonized [27, 28]. The sooner the duodenoscope is identified as the source of transmission, the more patients can be excluded from undergoing treatment with this contaminated duodenoscope and hence the risk of becoming infected.
Currently, some guidelines advise to perform microbiological surveillance periodically, but there is no consensus on the intervals. European guidelines [29] recommend surveillance at least every 3 months, Australian guidelines [30] recommend monthly surveillance, the American Society for Gastrointestinal Endoscopy does not urge periodic surveillance [31]. This case demonstrates that even more than once a month sampling can miss intermittent contamination and might create a false impression of safety. Only a culture-and-hold-method can fully prevent contaminated duodenoscopes from being used in ERCPs [32]. However, this requires an investment in extra duodenoscopes and in daily microbiological testing, which is not feasible in most endoscopy centers.
It is important to share information on DAI incidence and local reprocessing issues globally to stimulate awareness and the development of solutions. In a 2016 US senate report, it was mentioned that the FDA and CDC are only rarely actively notified by hospitals and endoscope manufacturers on these issues [33]. This has led to an official warning towards the three main duodenoscope manufacturers [34]. Timely reporting of incidents to the authorities, will sooner lead to actions protecting patients all over the world.
