Arjan Rauwers

Discussion and future perspectives

The worldwide surge of duodenoscope-associated outbreaks since the new millennium show that current reprocessing practices do not guarantee adequately decontaminated endoscopes. To prevent future outbreaks, identification of risk factors contributing to outbreaks and endoscope contamination is essential. The discussion of this thesis is divided into three themes. First, it describes the background of reprocessing of duodenoscopes and linear echoendoscopes (DLE), the multiple parties that are involved, and potential pathways leading to an outbreak. Secondly, it gains insight in the true size of the underlying problem of DLE contamination by showing the enduring high prevalence of DLE contamination with digestive tract bacteria in Dutch hospitals. In the third part of this thesis we show the results of a study of a cleaning test as a potential marker to check for residue to lower the contamination rate.

Part I: Duodenoscope-associated outbreaks
In Chapter 2 we discuss the developing story of duodenoscope-associated outbreaks in 2015 as a call for awareness for Dutch medical doctors of all specializations. Outbreaks reported by high impact media such as the LA times and Washington Post, and in 2018 in the Dutch newspaper AD, caused public disturbance, leading patients 1-3 to question ERCP safety. Contamination of endoscopes is a multidisciplinary issue involving manufacturers, regulatory agencies, hospital directors, gastroenterologists, medical microbiologists, infection prevention specialists, disinfection professionals and disinfections assistants, as well as medical specialists such as internists and surgeons whose patients also undergo ERCP procedures. Details of endoscope reprocessing were unknown by this audience and the scale of the problem was unclear. In the Netherlands, at that time, endoscope surveillance cultures were not mandatory as only process 4 control was used, which considers reprocessing to be adequate if performed according the manufacturer’s instructions for use (IFU). Outbreaks of MDRO were only reported by large tertiary centers which were able to detect and trace the outbreak using the distinct features of the MDRO. It is very likely that transmission of microorganisms occurs regularly while not recognized. Duodenoscope-associated infections (DAI) can also occur non-clustered, and a DAI with susceptible microorganisms can easily be mistaken 5 for an endogenous infection as a result of the ERCP procedure. In addition, DAI are 6 also under-reported and inadequately registered, leading to gross underestimation 9 of the prevalence. An adequate registration system of outbreaks and of transmission by endoscopes, including susceptible microorganisms, is still lacking. A recent review assessing published reports, identified up to almost 500 infected patients and over 30 7 deaths caused by contaminated duodenoscopes between 2008 and 2018, but this is still considered to be the tip of the iceberg. Positively, hospitals are currently more aware about the possibility of endoscope-associated infections, as shown by two recall actions 8, 9 by Dutch and Belgian hospitals in 2021 reported by local layman media. We also paid attention to current legislation which grants endoscopes with new designs market access based on their presumed similarity with previous models without the need for clinical trials. 10, 11 This led to patients being treated with equipment which design modifications affected patient safety. Furthermore, manufacturers failed to timely register adverse 12 events of their medical devices. This led to the criminal persecution of Olympus as 13 they did not timely filed the outbreak reported by the Erasmus MC in 2012. In reaction to the surge of outbreaks, the FDA proposed several solutions to reduce contamination rates including the development of duodenoscopes with disposable parts or single-use endoscopes. Although still not automatically required for new endoscope designs, we believe that studies should be conducted to prove their safety and added value.

In chapter 3 we performed a root cause analysis of contributing factors to an outbreak of multidrug-resistant Klebsiella pneumoniae in a Dutch tertiary academic hospital in 2015. This outbreak is exemplary that the current system of process control is not robust enough. We found that the outbreak was the result of a multitude of factors including the design of the duodenoscope, inadequate repairs, improper cleaning, miscommunication about cleaning protocols and lack of microbiological surveillance. 6 Recurrent outbreaks like these since the initial surge since 2008, show that DAI are an ongoing problem. Alertness by adequate surveillance, end control and recurring audits remains necessary. The attack rates (the number of infected or colonized cases/number of exposed persons) of 35% and 29% were similar to other outbreaks (12%-41%). 14-18 The outbreak period of 8 months show that contaminated endoscopes can remain undetected for months, just as other outbreaks had lengths of four or even up to twelve months. 16, 18-22 Following the outbreaks, half-yearly microbiological surveillance has been 23 introduced in the Netherlands. To reduce the risk of month-long outbreaks as a result of undetected use of contaminated endoscopes, more frequent surveillance could be considered. The FDA has ordered manufacturers to file new admissions tests to reevaluate endoscope designs, and manufacturers now suggest yearly servicing inspections. 24-27 28 As endoscope biopsy channels are frequently damaged, which can occur within a 29 few months of use, and duodenoscopes without indications for servicing can have critical abnormalities which may contribute to outbreaks, 21, 30 preventive inspections and maintenance must be considered. Clear communication between manufacturers and Independent Servicing Organizations as well as a transparent servicing market should enable hospitals to be sure that their repaired duodenoscopes is of similar quality as a brand-new one. Furthermore, to ensure reprocessing protocol adherence, manufacturers must communicate new reprocessing recommendations right away and hospitals must perform recurrent audits. Lastly, to improve the reliability of the defensive layers around reusable endoscopes and to avoid that internal manufacturer assessment goes unreported, 21, 31 after future outbreaks hospitals should report and publish their internal review.

Part II: Contamination of complex gastrointestinal endoscopes: prevalence and risk factors
The second part of this thesis focuses on the prevalence and risk factors of contamination of duodenoscopes and linear echoendoscopes (DLE). In chapter 4 we conducted the first nationwide study to assess the prevalence rate of duodenoscopes contaminated with digestive tract bacteria among all Dutch ERCP centers. We found that 22% of the duodenoscopes, originating from 26 (39%) centers were contaminated with any microorganism with ≥20 colony forming units (CFU)/20 mL (AM20). Moreover, we found that 15% of the Dutch duodenoscopes harbored microorganisms with gastrointestinal or oral origin (MGO). These results confirmed our hypothesis that patients undergoing ERCP are regularly being treated with contaminated equipment, and not detected as such. No difference was shown in contamination risk between the different duodenoscope types, which is in line with other studies and with reported outbreaks involving various duodenoscope types. 6, 32 Other studies found lower contamination rates, 30, 32 which could be explained by the continuous feedback of the post-procedure or everyday morning cultures of these studies. Another possibility is that we used a more sensitive sampling and culturing method, including a more sensitive contamination cut-off and a longer incubation time. This study showed that to minimize the risk of microbial transmission more stringent measures were required including microbiological surveillance.

In chapter 5 we present the results of the second nationwide prevalence study in which we assessed the contamination prevalence of duodenoscopes as well as linear echoendoscopes. The MGO contamination prevalence of 15% found with the PROCESS 33 2 study was similar to the first PROCESS study which was conducted two years earlier. This shows that not only patients undergoing ERCP but also EUS are being treated with contaminated equipment and that this risk had not lowered since PROCESS 1. The contamination prevalence was 8% if the definition by the Centers for Disease Control and Prevention for high-concern organisms was used. After publication of outbreaks in 2015, the FDA demanded post-market surveillance studies by Fujifilm, Olympus and Pentax which all three initially failed to conduct. 12, 34 Based on the first interim contamination rates of 3% for high concern organisms, the FDA stated that 12 reprocessing is not sufficient to avoid duodenoscope-associated infections. Current interim show a rate of 4.1%-6.1%, 34, 35 which is in line with results of the PROCESS studies. Studies assessing surveillance, 32, 35-37 culture and quarantine strategies, 30, 32, 38, 39 and disinfection interventions found lower rates. 40-42 These lower rates are promising and may be the result of continuous feedback and raised alertness by the studies’ culture results, 32, 39-41 but also less sensitive sampling and culture methods. Assessing the DLEs of both PROCESS studies showed that the age of older and younger endoscopes had similar contamination risks, which suggests that if old and heavily used DLEs are correctly maintained their contamination risk is similar to brand-new DLEs. Furthermore, contamination was independent of reprocessing characteristics. Probably manual cleaning is the most important factor, complicated by factors such as the complex endoscope design, endoscope damage and whether a biofilm has already formed. While initially the forceps elevator was thought to be the culprit as it was the source of contamination in several outbreaks, 18, 20, 21, 30 MGO rates of the biopsy channels flush (5%), suction channel flush (5%), and brush (8%) were higher than the forceps elevator (4%). Both PROCESS studies show that especially the endoscope channels harbor gastrointestinal microorganisms.

Part III: Long- and short-term solutions: a role for post-manual cleaning tests?
In the third and final part we investigated if use of post-cleaning adenosine triphosphate (ATP) tests lowers the number of DLE contaminated with gut flora. Multiple gastroenterology and regulatory agencies have stressed the need for easy control measures to check for adequacy of endoscope decontamination. 43, 44 Microbiological surveillance cultures are the current gold standard to assess if reprocessing was adequate, but have multiple downsides as they are labor intensive, expensive and give delayed feedback due to laboratory process time. The FDA and multiple endoscopy societies have suggested the ATP-test as an alternative as it is relatively cheap, easy to 45 perform, and it gives feedback within minutes. The test is a bioluminescence assay which emits ATP-dependent light measured in Relative Light Units (RLU), while using luciferase-catalyzed oxidation of luciferin. Although presence of ATP may indicate residual organic material which requires cleaning, it is unclear if use of ATP tests actually lowers the number of contaminated endoscopes. Therefore, we initiated the prospective single center before-and-after DETECT study (Duodenoscopes and linear echoendoscopes: Efficacy of ATP TEsts Compared to visual inspection).

In chapter 6 we present the results of the DETECT study. DLEs were ATP tested post-manual cleaning after 909 procedures. During the intervention period DLEs were recleaned if positive before proceeding to high-level disinfection (HLD). DLEs underwent microbiological sampling after HLD. Introduction of ATP tests did not reduce the gut flora contamination rate and did not stop or prevent the use of endoscopes contaminated with gut flora. Although contamination with gut flora was less likely to occur during the intervention period (OR 0.32; 95% CI 0.12-0.85), the absolute gut flora contamination rate was higher (16%; n=67 vs. 21%; n=102). The lower odds on gut flora contamination in the intervention period were based on two duodenoscopes. During the control period, these duodenoscopes had multiple episodes of ongoing contamination with the same microorganism. The episodes were ended by quarantining and repairs before the start of the intervention period and were thus not terminated by ATP testing.

This is the first controlled and adequately powered study with blinded study cultures, which confirms the results of previous smaller studies that post-cleaning ATP tests are not effective. 46, 47 False negative tests may provide a false sense of security, as during the intervention period the number of positive ATP-tests were lower while gut flora rates remained high. The rates in this study, which was blinded for study culture outcomes, were far higher compared to other large-scale open design studies with post-procedure or daily surveillance, 30, 32, 38 but may be more representative for clinical practice. Especially the rates for Pentax duodenoscopes were high, potentially the result of biofilm formation in the channels. The high contamination rates and undetected contamination episodes show that monthly surveillance cultures were inadequate. Until the risk of transmission via contaminated equipment is eliminated or reliable quality checks are introduced, more frequent microbiological surveillance is indicated.

The results of the DETECT study showed a discrepancy between ATP-positive test results and post-HLD presence of gut flora. In chapter 7 we assess the predictive value of post-cleaning ATP tests for the presence of microorganisms after HLD, using data from the DETECT study. We compared the RLU levels of ATP tests performed after manual cleaning with growth of any microorganism in post-HLD cultures. Nor the RLU cut-off according to the instructions for use, nor any other cut-off was found to be of clinical value: all the ROC curves per sample site, DLE type and microorganism type had an area under the curve of <0.6, with a low sensitivity and in most sample sites also a low specificity. This showed that post-cleaning ATP tests could not predict the absence or presence of microorganisms after HLD. This is in line with earlier pilot studies which also did not found a correlation between post-cleaning ATP tests and post-HLD cultures. 47, 48 The difference between sites with high ATP test results and negative cultures could be explained by a reduction of debris by HLD. The ATP values can also reflect presence of other debris containing ATP but no microorganisms such as blood. The combination of negative ATP tests and positive cultures could be the result of microorganisms in low concentrations, unable for the ATP test to detect. Another explanation might be the sampling method of the channel which is most often contaminated. The ATP flush of the channel is strongly diluted, while filtration of the culture flush contributes to detection of all present microorganisms. The low diagnostic accuracy of post-cleaning ATP levels for presence of post-HLD microorganisms means that ATP-tests have no clinical value in improving the outcome of DLE reprocessing. Future perspectives and recommendations In chapter 8 we look forward to the future, discussing opportunities and directions which may reduce and potentially eliminate endoscope-associated infections. In the coming years major improvements in reprocessing techniques and endoscope design such as endoscopes with disposable elements or even single-use endoscopes are expected. However, the majority of the hospitals will continue to use reusable endoscopes in at least the coming decade. Therefore, process control as well as regular audits including training of cleaning personnel to follow the IFU to the letter and 49 providing them with ample time to complete all reprocessing steps remain essential. Also microbiological surveillance remains the gold standard and the bare minimum to prevent ongoing transmission of microorganisms, as no bioburden assay has proven useful yet. Endoscopy societies should reach consensus on the sampling and culturing methods, as well as on the optimal culture frequency. The other short-term solutions include clear communication between all parties involved, endoscope risk evaluation and improvement of reprocessing methods. As a result of FDA safety communications, 6 a critical evaluation by the US senate, and updated multisociety guidelines 45, 50 awareness on the subject has improved, leading to better communication between gastroenterologists, medical microbiologists and infection prevention professionals. In the Netherlands in particular, awareness has improved because of participation of 51 all Dutch hospitals in the two PROCESS studies, and introduction of a new Dutch 23 microbiological surveillance guideline in 2018. Members of our research group contributed to this guideline which was supported by all involved Dutch societies. Furthermore, transparent communication, reporting and evaluation of adverse events (i.e. outbreaks, device failures and reprocessing risks) as well as post-market follow-up of endoscope designs remains paramount. While initially the forceps elevator was seen as the culprit for contamination, evidence now shows the entire endoscope and in particular the channels may harbor 51 microorganisms, independent of physical age or usage. During outbreaks it was shown that normal-functioning duodenoscopes had critical abnormalities, 21, 30, 31 which may 52 affect the contamination risk. Therefore manufacturers should consider usage-based inspections in combination with an endoscope specific log file including previous repairs and culture results. This facilitates recognition of endoscope design specific flaws and the need for preventive maintenance. To reduce the contamination risk of reusable endoscopes, several reprocessing measures in addition to the manufacturers’ IFU have been proposed including ethylene oxide (EtO) sterilization and repeat HLD. They show promising results, but are costly and do not provide zero contamination rates. The current method of reprocessing including HLD has a very small margin of safety leaving no room for error. 53-56 As ERCP procedures are increasingly more invasive and often natural mucosal barriers are breached, new reprocessing methods should be developed which have a larger margin of safety than currently required for reusable duodenoscopes. Ultimately, the risk of contamination must be eliminated by radically redesigned endoscopes which can be sterilized or single-use endoscopes. Currently redesigned duodenoscopes are introduced with disposable elements such as disposable protection caps, 57, 58 a disposable forceps elevator or with a sterilizable removable elevator mechanism. 59, 60 These endoscopes will not hold the solution, as the other parts remain unchanged and the endoscope simply cannot be sterilized. Furthermore, these endoscope designs are not peer-reviewed. An important lesson from the failed introduction of the TJF-Q180V duodenoscope, is that successive changes in endoscope design can lead to safety risks. Therefore, the current situation of market authorization of new endoscopes without clinical tests if the modified design is technical similar to a previous approved design, as judged by the manufacturer, needs to change. Single-use endoscopes show promise in overall technical performance and safety 61 profile comparable to reusable duodenoscopes, but implementation will depend on their cost-effectiveness. The clinical question which is still unanswered is how often transmission of microorganisms via endoscopes occurs and how many times it leads to a clinically relevant infection or colonization of the patient. A first retrospective revision by Kwakman et al., estimated a risk of duodenoscope-associated infections 62 63 (DAI) of 0.01%: at least 180 times higher than previous estimates. To assess the true incidence, prevalence, patient burden and financial costs of DAI, one would ideally conduct a prospective multicenter study blinded for study cultures and including (pre-endoscopy) patient cultures. Newly designed (single-use) endoscopes will not be a widespread and readily available solution to eliminate the transmission risk. Therefore the application and decontamination of reusable DLE destined to be in operation for many years to come, must be improved in order to lower contamination rates.