- Email: [email protected]
Randomized Comparison of 3 High-Level Disinfection and Sterilization Procedures for Duodenoscopes
Accepted Manuscript Randomized Comparison of 3 High-level Disinfection and Sterilization Procedures for Duodenoscopes Graham M. Snyder, MD SM, Sharon B. Wright, MD MPH, Anne Smithey, Meir Mizrahi, MD, Michelle Sheppard, RN MSN DNPc, Elizabeth B. Hirsch, PharmD, Ram Chuttani, MD, Riley Heroux, David S. Yassa, MD MPH, Lovisa B. Olafsdottir, MD, Roger B. Davis, ScD, Anastasiou Jiannis, MD, Vijay Bapat, MD, Kiran Bidari, MD, Douglas K. Pleskow, MD, Daniel Leffler, MD MS, Benjamin Lane, BS, Alice Chen, Howard S. Gold, MD, Anthony Bartley, MD, Aleah D. King, RN BSN, Mandeep S. Sawhney, MBBS MS PII: DOI: Reference:
S0016-5085(17)35869-9 10.1053/j.gastro.2017.06.052 YGAST 61271
To appear in: Gastroenterology Accepted Date: 27 June 2017 Please cite this article as: Snyder GM, Wright SB, Smithey A, Mizrahi M, Sheppard M, Hirsch EB, Chuttani R, Heroux R, Yassa DS, Olafsdottir LB, Davis RB, Jiannis A, Bapat V, Bidari K, Pleskow DK, Leffler D, Lane B, Chen A, Gold HS, Bartley A, King AD, Sawhney MS, Randomized Comparison of 3 High-level Disinfection and Sterilization Procedures for Duodenoscopes, Gastroenterology (2017), doi: 10.1053/j.gastro.2017.06.052. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
The DISINFECTS trial is a prospective randomized trial investigating three methods of reprocessing duodenoscopes: standard high-level disinfection with ortho-phthalaldehyde disinfectant, standard highlevel disinfection with a repeated (double) cycle of disinfectant exposure, and standard high-level disinfection followed by ethylene oxide gas sterilization. No significant difference was noted between these reprocessing methods with regard to proportion of duodenoscopes with MDRO contamination, or demonstrating ≥10 CFU or >0 CFU growth of any aerobic bacteria.
ACCEPTED MANUSCRIPT
Title: Randomized Comparison of 3 High-level Disinfection and Sterilization
Short Title: Duodenoscope Disinfection Trial
RI PT
Procedures for Duodenoscopes
Authors: Graham M. Snyder MD SM1,2, Sharon B. Wright MD MPH 1,2, Anne
SC
Smithey1,4, Meir Mizrahi MD3, Michelle Sheppard RN MSN DNPc3, Elizabeth B. Hirsch PharmD4, Ram Chuttani MD2,3, Riley Heroux1,4, David S. Yassa MD MPH1,2, Lovisa B.
M AN U
Olafsdottir MD1, Roger B. Davis ScD5, Anastasiou Jiannis MD3, Vijay Bapat MD3, Kiran Bidari MD3, Douglas K. Pleskow MD2,3, Daniel Leffler MD MS2,3, Benjamin Lane BS1, Alice Chen1,4, Howard S Gold MD1,2, Anthony Bartley MD3, Aleah D. King RN BSN1, Mandeep S. Sawhney MBBS MS2,3
TE D
1. Division of Infection Control/Hospital Epidemiology, Beth Israel Deaconess Medical Center, Boston, Massachusetts
2. Harvard Medical School, Boston, Massachusetts
Massachusetts
EP
3. Division of Gastroenterology, Beth Israel Deaconess Medical Center, Boston,
AC C
4. Department of Pharmacy and Health Systems Sciences, Northeastern University, Boston, Massachusetts
5. Division of General Medicine and Primary Care, Beth Israel Deaconess Medical Center, Boston, Massachusetts
Grant support:
1
ACCEPTED MANUSCRIPT
This work was funded in part by an investigator-initiated grant from the American Society for Gastrointestinal Endoscopy (G.M.S., M.S.S.) and from Beth Israel Deaconess Medical Center, Boston, Massachusetts. Study sponsors did not play any
RI PT
role in the study design, collection, analysis, and interpretation of data.
Abbreviations
SC
CFU, colony-forming units
HLD, high-level disinfection
M AN U
dHLD, double high-level disinfection
HLD/ETO, standard high-level disinfection followed by ethylene oxide gas sterilization MDRGN, multidrug-resistant Gram-negative pathogens MDRO, multidrug resistant organisms
TE D
MRSA, methicillin-resistant Staphylococcus aureus sHLD, standard high-level disinfection
EP
VRE, vancomycin-resistant enterococci
Main Outcomes and Measures: The primary outcome was the proportion of
AC C
duodenoscopes with an elevator mechanism or working channel culture showing ≥1 MDRO; secondary outcomes included the frequency of duodenoscope contamination with >0 and ≥10 colony-forming units (CFU)
Corresponding Author: Mandeep Sawhney, MBBS MS
2
ACCEPTED MANUSCRIPT
330 Brookline Ave, Rabb-Rose 133 Boston, MA 02215, USA Phone: (617) 667-1088
RI PT
Fax: (617) 667-1171 Email: [email protected]
SC
Disclosures:
M AN U
Conflicts of Interest. All authors report no other potential conflicts of interest to disclose.
Author Contributions:
Conception/design: GMS, SBW, AS, MM, EBH, DSY, RBD, DL, HSG, AB, ADK, MS Data collection: AS, MM, MS, RC, RH, LBO, AJ, VB, KB, DKP, AB, MS
TE D
Data analysis: GMS, AS, EBH, RH, LBO, RBD, BL, AC Manuscript preparation: GMS, SBW, MSS, RBD
EP
Approval of final version of the manuscript: all authors
ACKNOWLEDGEMENTS
AC C
Portions of this data have been presented at the Society for Healthcare Epidemiology of America Spring 2016 Conference and at Digestive Disease Week 2017. This work was conducted with support from Harvard Catalyst (NIH Award UL1 TR001102).
The authors would like to extend our appreciation to James Kirby, M.D. for his advice on the design of the study, Nancy Doraiswami and Maxine Hainstock for facilitating
3
ACCEPTED MANUSCRIPT
additional reprocessing required for the study, and Christopher Rowley, M.D. and Alan Moss, M.D. for their advising as independent members of the Safety Review
AC C
EP
TE D
M AN U
SC
Endoscopy Unit for their work and support on the project.
RI PT
Committee. Additionally, we would like to thank the staff of the Stoneman 4 Advanced
4
ACCEPTED MANUSCRIPT
ABSTRACT Background and Aims: Duodenoscopes have been implicated in the transmission of multi-drug resistant bacteria (MDRO). We compared the frequency of duodenoscope
RI PT
contamination with MDRO or any other bacteria after disinfection or sterilization by 3 different methods.
SC
Methods: We performed a single-center prospective randomized study in which
duodenoscopes were randomly reprocessed by standard high-level disinfection (sHLD),
M AN U
double high-level disinfection (dHLD), or standard high-level disinfection followed by ethylene oxide gas sterilization (HLD/ETO). Samples were collected from the elevator mechanism and working channel of each duodenoscope and cultured before use. The primary outcome was the proportion of duodenoscopes with an elevator mechanism or
TE D
working channel culture showing 1 or more MDRO; secondary outcomes included the frequency of duodenoscope contamination with more than 0 and 10 or more colony-
EP
forming units (CFU) of aerobic bacterial growth on either sampling location.
Results: After 3 months of enrollment, the study was closed due to the futility—we did
AC C
not observe sufficient events to evaluate the primary outcome. Among 541 duodenoscope culture events, 516 were included in the final analysis. No duodenoscope culture in any group was positive for MDRO. Bacterial growth of more than 0 CFU was noted in 16.1% duodenoscopes in the sHLD group, 16.0% in the dHLD group, and 22.5% in the HLD/ETO group (P=.21). Bacterial growth or 10 or more CFU was noted in 2.3% of duodenoscopes in sHLD group, 4.1% in the dHLD group, and
5
ACCEPTED MANUSCRIPT
4.2% in the HLD/ETO group (P=.36). MRDOs were cultured from 3.2% of pre-procedure rectal swabs and 2.5% of duodenal aspirates.
RI PT
Conclusions: In a comparison of duodenoscopes reprocessed by sHLD, dHLD, or
HLD/ETO, we found no significant differences between groups for MDRO or bacteria contamination. Enhanced disinfection methods (dHLD or HLD/ETO) did not provide
SC
additional protection against contamination. However, insufficient events occurred to
M AN U
assess our primary study end-point. ClinicalTrials.gov no: NCT02611648
AC C
EP
TE D
KEY WORDS: DISINFECTS study; surveillance; endoscopy; non-outbreak setting
6
ACCEPTED MANUSCRIPT
INTRODUCTION Recent outbreaks of infections due to carbapenem-resistant Enterobacteriaceae and other multidrug-resistant organisms (MDRO) have been attributed to contaminated
RI PT
duodenoscopes, despite reported compliance with manufacturer recommendations for reprocessing.1-6 The reasons for duodenoscope contamination when reprocessing is conducted appropriately are not fully understood, and may relate to biofilm formation
SC
within the duodenoscope.
An advisory panel convened by the Food and Drug Administration agreed that
M AN U
current standards for duodenoscope reprocessing did not provide a reasonable level of safety and effectiveness.7 Experts have suggested the adoption of additional measures like double reprocessing cycles and ethylene oxide sterilization to supplement current standards.8-10 These additional measures are expensive, time consuming and not
TE D
readily available in all institutions. Furthermore, the effectiveness of these additional measures has never been systematically studied in a non-outbreak setting. Studies comparing different methods of duodenoscope reprocessing are therefore needed if we
EP
are to identify the optimal method of duodenoscope reprocessing. In the DISINFECTS trial (Duodenoscope Infection Surveillance IN Functioning
AC C
automated Endoscope reprocessors in Conjunction with eThlyene oxide Sterilization), we aimed to compare the frequency of duodenoscope contamination with MDRO and with any bacterial contamination after reprocessing using three methods of high-level disinfection or sterilization.
METHODS
7
ACCEPTED MANUSCRIPT
Study design and setting We conducted a prospective randomized trial investigating three methods of reprocessing duodenoscopes: standard high-level disinfection with ortho-
RI PT
phthalaldehyde disinfectant (sHLD), standard high-level disinfection with a repeated (double) cycle of disinfectant exposure (dHLD), and standard high-level disinfection followed by ethylene oxide gas sterilization (HLD/ETO). The study was approved by the
SC
Institutional Review Board at our institution, and was registered at clinicaltrials.gov
(NCT02611648). All authors had access to study data and reviewed and approved the
M AN U
final manuscript.
During the study period, all duodenoscopes selected for patient use were cultured prior to the procedure. Patients undergoing an endoscopy procedure for which the use of a duodenoscope was anticipated were offered participation in the study. Prior
TE D
to patient enrollment, 18 duodenoscopes were assigned to one of three study arms. Reprocessing time for duodenoscopes assigned to the HLD/ETO arm was estimated to be substantially longer than the other arms; we therefore assigned eight
EP
duodenoscopes to the HLD/ETO arm, and five each to the sHLD and dHLD study arm. Duodenoscope assignment to a study arm was retained for the duration of the study. A
AC C
permuted block randomization key specifying study arm was randomly generated with a block size of three (comprising one from each study arm). Endoscopy procedures were performed according to standard clinical practice. The attending endoscopist and patient were blinded to duodenoscope study arm assignment.
8
ACCEPTED MANUSCRIPT
Patient history of MDRO carriage was obtained from Infection Control databases at the study institution. At the time of the procedure, a rectal swab and duodenal aspirate sample were obtained to identify carriage with MDRO.
RI PT
Study Procedures Devices and reprocessing
Duodenoscope reprocessing during the study entailed a manual wipe of the
SC
exterior of the duodenoscope with enzymatic solution (EmPower, Metrex) immediately following the procedure, manual reprocessing within one hour of the procedure
M AN U
including the use of device-specific elevator mechanism and working channel brushes, and completion of reprocessing using automated endoscope reprocessors (System 83 Plus 9, Custom Ultrasonics) with ortho-phthalaldehyde disinfectant (MetriCide OPA Plus, Metrex) followed by ethanol flush and vertical hanging for drying in a non-
TE D
ventilated cabinet. Duodenoscopes assigned to the dHLD arm, underwent reprocessing similar to the sHLD arm, except 2 re-processing cycles of automated endoscope reprocessor with ortho-phthalaldehyde disinfectant were used. For duodenoscopes
EP
assigned to the HLD/ETO trial arm, ethylene oxide gas sterilization was performed following these reprocessing steps (Steri-Vac Sterilizer/Aerator, 3M). Prior to study
AC C
initiation, our endoscopy unit utilized seven duodenoscopes (Olympus, model TJFQ180V) all of which were acquired in 2012; in order to accommodate additional reprocessing time associated with our study, an additional 11 new duodenoscopes of the same manufacturer and model were acquired in 2015. The 2012 and 2015 duodenoscopes were proportionally distributed to study arms. Duodenoscope and patient sample collection and microbiological methods
9
ACCEPTED MANUSCRIPT
The duodenoscope culturing process consisted of two samples obtained using stringent aseptic technique: 1) a sterile dry flocked swab sample of the elevator mechanism, and 2) a sterile saline sample of the working channel using the “flush-
RI PT
brush-flush” technique.11 The flocked swab elevator mechanism sample was placed into transport media (ESwab with liquid Amies media, Copan Diagnostics). The working channel sample was obtained by flushing sterile water through the working channel,
SC
passing a sterile device-specific brush (Olympus) through the length of the working channel and manually agitated in the captured sterile water after removal, and then
M AN U
flushing another aliquot of sterile water through the working channel and captured with the initial flush specimen. A standardized volume of sample was added to phosphatebuffered saline for the transport of working channel samples.12 Duodenoscopes were sampled in a dedicated location. Duodenoscopes cultured but not used within one
TE D
calendar day following culture were reprocessed without patient use. Duodenoscope sample cultures were processed for two outcomes: growth of any aerobic bacteria (colony-forming units, CFU), and the presence of MDRO. MDRO were defined as
EP
methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), or multidrug-resistant Gram-negative pathogens (MDRGN).
AC C
Among patients providing informed consent for the patient microbiologic
component of the study, a rectal swab (CultureSwab Plus Amies Gel without Charcoal, BD) and duodenal aspirate were obtained at the time of the procedure. Patients were given the opportunity to opt-out of the rectal swab collection. Patient samples were processed for the presence of MDRO. The detailed duodenoscope and patient sample collection and microbiologic methods can be found in the Supplementary Material.
10
ACCEPTED MANUSCRIPT
Data collection and safety monitoring Patient- and procedure-related information was collected by study investigators using a structured intake form developed for the study. An independent observer
RI PT
blinded to trial arm assignment reviewed duodenoscope culture results each weekday to identify duodenoscopes that met pre-specified criteria for immediate review by the Safety Monitoring Committee. Blinded data were reviewed in relation to potentially
SC
defective duodenoscopes and loss of intervention arm equipoise. Statistical analysis
M AN U
The final analysis included all procedures in which a randomly assigned duodenoscope was used on a patient and both the elevator mechanism and working channel were sampled. The primary outcome of the study is the proportion of duodenoscopes in each of the three disinfection arms contaminated with ≥1 MDRO on
TE D
either the elevator mechanism or working channel. Secondary outcomes included the proportion contaminated with ≥10 CFU and >0 CFU of any aerobic bacteria on either the elevator mechanism or working channel in the three disinfection arms, and the
EP
frequency of patient MDRO acquisition after exposure to a duodenoscope with MDRO contamination. Differences in proportions of events among study arms were calculated
AC C
using Fisher’s exact test. We used a group sequential design with a Haybittle stopping rule to allow monthly interim analyses while preserving a study-wide 5% level of significance for the primary outcome. Blinded results were presented to the Safety Monitoring Committee, which made recommendations concerning continuation of the study.
11
ACCEPTED MANUSCRIPT
To explore factors associated with duodenoscope contamination with either ≥1 MDRO, any growth (>0 CFU) or significant contamination (≥10 CFU), on either the elevator mechanism or working channel, we fit Poisson regression models with trial arm
RI PT
as the factor of interest to obtain estimates of relative risk between standard disinfection and each of the more intensive methods. Models were fit using generalized estimating equations methods with an autoregressive correlation structure to account for
SC
correlation among observations from the same duodenoscope. To address for
confounding by factors other than trial arm that may be associated with duodenoscope
M AN U
contamination, covariates included in an adjusted model included duration of preceding procedure (in hours), use of sphincterotomy and stent removal during the preceding procedure, and time elapsed between preceding procedure and repeat culture (in days). We calculated a sample size assuming 5% contamination in the standard
TE D
practice trial arm (sHLD) and 0.5% contamination in the intensive intervention trial arm (HLD/ETO). We set the level of significance to 5% and power to 70%. With adjustments for the 3-arm design (using the least favorable distribution assumption) and the group
AC C
RESULTS
EP
sequential design, a sample size of 206 per trial arm was required.
After two months of study enrollment, the Safety Review Committee
recommended that patient microbiologic data collection be stopped due to very low probability of observing patient MDRO acquisition events, and after the third month of enrollment recommended that the study close due to the futility of observing sufficient events to evaluate the primary outcome.
12
ACCEPTED MANUSCRIPT
Duodenoscope contamination Among the 541 study observations, 531 (98.2%) duodenoscopes were cultured prior to the procedure, 517 (95.6%) duodenoscopes were used following appropriate
RI PT
randomization, and 516 (95.4%) were cultured and appropriately randomized prior to selection for a patient procedure. These 516 observations intended for patient use were included in the primary analysis including 174 (33.7%) assigned to the sHLD arm, 169
SC
(32.8%) to the dHLD arm, and 173 (33.5%) to the HLD/ETO arm.
No duodenoscope cultures grew ≥1 MDRO from either the elevator mechanism
M AN U
or working channel. Thus there was no difference between the three study arms with regard to the primary study outcome (Table 1). Twenty (3.9%) duodenoscope cultures demonstrated growth of ≥10 CFU in either the elevator mechanism or working channel, including 11 (2.0%) on the elevator mechanism only, eight (1.5%) on the working
TE D
channel, and one (0.2%) on both the elevator mechanism and working channel. There was no difference between the three study arms with regard to >0 CFU or ≥10 CFU growth of any aerobic bacteria (Table 1). Figure 1 demonstrates the frequency
EP
of no growth (0 CFU), low quantity of growth (1-9 CFU) and significant quantity of growth (≥10 CFU) in each of the study arms. Figure 2 demonstrates the pattern of
AC C
positive cultures on individual duodenoscopes throughout the study period. No duodenoscopes met pre-specified criteria for persistent duodenoscope contamination. Duodenoscope contamination regression modeling A regression model demonstrated no association between study arm and
frequency of duodenoscope contamination when >0 CFU bacterial growth was used as the outcome variable (Table 2). No significant association between the enhanced study
13
ACCEPTED MANUSCRIPT
arms and contamination was noted even after adjusting for duration of preceding procedure, sphincterotomy, stent removal, and time elapsed between preceding procedure and repeat culture. The regression model with the outcome of ≥10 CFU
RI PT
growth modeled without covariates demonstrated an approximately 2-fold higher
likelihood of duodenoscope contamination among the dHLD and HLD/ETO arms compared to the sHLD arm (Table 2).
SC
Duodenoscope exposure to patient carriage of MDRO
For 43 of the 541 (7.9%) study observations the duodenoscope did not have
M AN U
patient contact as either >1 calendar day had passed between the time of culture and demand for duodenoscope use (n=16), or duodenoscopes were obtained for use but did not subsequently have patient contact (n=27). The remaining 498 procedures were performed on 389 patients, including three (0.8%) patients who underwent four
TE D
procedures, 15 (3.9%) patients who underwent three procedures, and 70 (18.0%) patients who underwent two procedures. Among the 389 patients undergoing at least one procedure during the study, 26 patients had a known history of carriage with ≥1
EP
MDRO prior to study participation. This included MRSA, 15 (3.9%); VRE, 12 (3.1%); and MDRGN, 9 (2.3%). Three patients had a history of both MRSA and VRE, one
AC C
patient with both VRE and MDRGN, and three patients with a history of MRSA, VRE, and MDRGN carriage.
For 343 of the 498 (68.9%) procedures patients were offered participation in the
patient microbiologic component of the study. Among these 343 procedures, 260 (75.8%) patients provided full consent (236 of 260 patients) or opted out of collection of rectal samples (24 of 260 patients). Twenty-four (7.0%) patients declined to participate,
14
ACCEPTED MANUSCRIPT
and for 59 (17.2%) patients consent was not sought due to procedure urgency. The remaining 155 of 498 (31.1%) patients underwent procedures after discontinuation of the patient microbiologic component of the study. The demographic characteristics of all
RI PT
study patients including patients undergoing procedures with duodenoscopes from each study arm are presented in Table 3.
Among the 260 consenting patients, rectal sampling for MDRO was obtained
SC
among 189 (72.7%) and duodenal aspirates among 202 (77.7%). On rectal swabs,
MRSA, VRE, and MDRGN were identified among one (0.5%), three (1.6%), and three
M AN U
(1.6%) of samples, respectively; a total of six samples (3.2%) had ≥1 MDRO including one patient with co-carriage of VRE and MDRGN. Among duodenal aspirates, MRSA, VRE, and MDRGN were identified among two (1.0%), two (1.0%), and one (0.5%) of
DISCUSSION
TE D
samples, respectively, including one MDRO for each of these five (2.5%) samples.
We report the results of the first prospective randomized trial comparing sHLD, dHLD
EP
and HLD/ETO for reprocessing duodenoscopes. In a non-outbreak setting we found no significant difference between these methods with regard to proportion of
AC C
duodenoscopes with MDRO contamination, or demonstrating ≥10 CFU or >0 CFU growth of any aerobic bacteria. In our study, MDRO were not recovered from any duodenoscope culture, and this limited our ability to detect differences between study arms with regard to the primary study end-point.
15
ACCEPTED MANUSCRIPT
The clinical events that prompted this trial were the reported outbreaks of patients acquiring MRDOs from contaminated duodenoscopes (1-6). We are unaware of any recent published reports of non-MDRO related duodenoscope acquired infections. We
RI PT
therefore focused on MDROs as a clinically relevant primary study end-point. When we designed the study in 2015 there were no published data on the proportion of
duodenoscopes contaminated with MRDOs. Using a conservative approach based on
SC
studies characterizing high-pathogenicity organisms, we powered our study for 5% of duodenoscopes in the sHLD arm showing contamination with MRDOs. Our study, in a
M AN U
tertiary care referral center, found that with vigilance in reprocessing MDRO contamination did not occur among any 516 duodenoscope culture events. Similar results have since been reported by Brandabur et al, who found 0 of 4032 duodenoscope cultures positive for MRDOs (16). In our study, 3.2% of rectal swabs
TE D
and 2.5 % of duodenal aspirates showed colonization by MRDO. These findings suggest that in the non-outbreak setting, sHLD may be able to adequately disinfect duodenoscopes even when endoscopes are used for procedures on a proportion of
EP
patients with MDRO colonization.
AC C
Duodenoscope contamination, regardless of the pathogenicity of the organism or moment of contamination, reflects a possible defect in reprocessing, handling and storage, or the device itself that may place the patient at risk of transmitted MDRO. To account for possible contamination from the time the duodenoscope was withdrawn from a patient until it was re-inserted into the next patient, we determined that significant bacterial contamination immediately prior to use would be more representative of this
16
ACCEPTED MANUSCRIPT
entire process than simply the presence or absence of pathogenic bacteria. We therefore used growth of ≥10 CFU as recommended by the Centers for Disease Control and Prevention to indicate potentially significant contamination (13). We also analyzed
RI PT
as an outcome any growth (> 0 CFU) from duodenoscope samples. Duodenoscope sampling was conducted in a dedicated space, and was performed by trained study personnel based using a technique based upon CDC sampling guidelines (13). Culture
SC
and lab techniques were validated using a standardized inoculum in the usual fashion. Additionally, we have previously demonstrated that our method of specimen storage
M AN U
and transport is unlikely to affect our findings (19). Following sHLD, the proportion of duodenoscopes with >0 CFU growth in our study was similar to growth rates reported by others, and thus provides a measure of validity to our culture methods (14-16). Considering only duodenoscopes assigned to the sHLD arm in our study, 16.1% (95%
TE D
CI, 11.2% – 22.1%) showed microbial growth > 0 CFU. Ross et al. reported that 13.1% of cultures collected from duodenoscopes after sHLD were positive for any bacterial growth (14). Brandabur et al reported on curvilinear echoendoscopes and
EP
duodenoscope cultures following sHLD and found up to 14.3% of Olympus endoscopes showed microbial growth (16). The average growth rate for any organism was 8.4% in
AC C
that study. There was very large variation amongst study sites with culture positive rates ranging from 0% to 35%.
At the Gastroenterology and Urology Devices Panel of the Medical Devices Advisory Committee meeting convened by the FDA in 2015, experts strongly urged the FDA to mandate that all duodenoscopes, and preferably all gastrointestinal endoscopes, be
17
ACCEPTED MANUSCRIPT
sterilized by 2018 (20). The FDA has not mandated ethylene oxide sterilization of duodenoscopes, but the use of ethylene oxide sterilization has been strongly proposed (21). The results of our trial do not support the use of ethylene oxide sterilization in a
RI PT
non-outbreak setting. The absence of any duodenooscope cultures showing MDROs lowered the statistical power of our study to detect difference between HLD/ETO and other study arms with regard to our primary end-point. However, the HLD/ETO arm in
SC
fact appeared inferior to sHLD with regard to our secondary end-point (proportion of duodenscope with ≥10 CFU growth). It is highly unlikely that even if we had continued
M AN U
to recruit subjects into the study, this trend would have reversed in the opposite direction to favor the HLD/ETO arm. In the HLD/ETO arm of our study, we found 22.5% of duodenoscopes showed microbial growth >0 CFU, of which 4.2 % showed microbial growth ≥10 CFU; compared with the sHLD arm in which only 16.1% of duodenoscopes
TE D
showed microbial growth >0 CFU, of which 2.3 % showed microbial growth ≥10 CFU. In our regression modeling, duodenoscopes assigned to the HLD/ETO arm in our study were almost twice as likely to show ≥10 CFU microbial growth when compared with the
EP
sHLD arm, and this difference was statistically significant (Table 2). There was also a non-significant increase in the probability of any microbial growth (>0 CFU) in the
AC C
HLD/ETO arm. Similarly, the dHLD arm also appeared have a high proportion of duodenoscopes with > 10 CFU microbial growth when compared with the sHLD arm (4.1% versus 2.3%). Apart from our study, we found only one other study that reported on duodenoscope cultures following ethylene oxide sterilization. Following an outbreak of E. coli at their institution, Naryzhny et al. instituted a policy of HLD/ETO for all duodenoscopes and found that 1.2% remained contaminated with high risk organisms
18
ACCEPTED MANUSCRIPT
(17). This contamination rate is similar to contamination rates seen with sHLD, again suggesting that HLD/ETO may not be superior to sHLD for disinfecting duodenoscopes.
RI PT
Some experts have suggested that ethylene oxide sterilization may increase likelihood of damage to duodenoscopes, and we hypothesize that this may have played a role in the increased duodenoscope contamination noted in the HLD/ETO arm (28). We did not
SC
find an association between the risk of duodenoscope contamination and duration of preceding procedures or performance of stent removals to suggest that instrumentation
M AN U
of the biopsy channel may have played a role in duodenoscope damage and hence contamination. Ross et al found that 3 of 4 duodenoscope that were contaminated with drug-resistant E coli were damaged and needed critical repairs (14). Interestingly, those duodenoscopes did not demonstrate any functional problems and those defects
TE D
were not recognized until the duodenoscopes were returned to the manufacturer. It is not known if dHLD or HLD/ETO sterilization can reliably disinfect structurally defective duodenoscopes. It is also possible that increased contamination seen in the HLD/ETO
EP
arm may be related to the increased handling of the duodenoscopes in that arm. However, the proportion of duodenoscopes with > 10 CFU growth was almost the same
AC C
between the HLD/ETO and the dHLD arm (4.2% versus 4.1%). No extra handling of duodenoscopes is likely to have taken place in the dHLD arm as the dHLD arm was similar to the sHLD arm except that the duodenoscopes underwent an extra automated reprocessing cycle. Differences in duodenoscope storage time between the study groups are also an unlikely explanation for these results. Recently our group has shown that there is no associated between duodenoscope storage time and risk of
19
ACCEPTED MANUSCRIPT
bacterial contamination (22). Ethylene oxide sterilization has several other drawbacks including a substantial increase in cost. HLD using an automated endoscope reprocessor can be performed in less than one hour. Ethylene oxide sterilization
RI PT
increases the duodenoscope processing time to 16 - 24 hours, thus requiring the
purchase of additional duodenoscopes. Naryzhny et al. estimated that using ethylene oxide sterilization increased their institutional expenditure by $ 93,563 over an 18
SC
months period. This estimate did not include additional personnel and increased
duodenoscope maintenance costs (17). In a cost-effective analysis, Almario et al.
M AN U
estimated that the use of ethylene oxide sterilization would increase the cost of each ERCP by $ 1043 (18). Considering that more than 600,000 ERCPs are performed in the United States annually, this would add more than half billion dollars to health care expenditure. Based upon these factors and our study results, we do not support the
TE D
routine use of ethylene oxide sterilization for duodenoscope reprocessing.
Our study has significant implications for further research in this field. With careful
EP
sHLD the proportion of duodenoscopes contaminated with MRDOs, as shown by our study, is likely to be extremely small. Therefore future studies conducted in non-
AC C
outbreak setting should not use MRDOs as the primary outcome, and instead use a surrogate marker of duodenoscope contamination. Proportion of duodenoscopes contaminated with pathogenic bacteria, proportion of duodenoscopes demonstrating significant aerobic growth (≥10 CFU), or proportion of duodenoscope with any aerobic bacteria growth (>0 CFU) are options for such a surrogate end-point. The optimal surrogate end-point should reflect duodenoscope reprocessing, handling and storage,
20
ACCEPTED MANUSCRIPT
and the possible structural damage to the duodenoscope itself. This is especially important when enhanced methods of disinfection are being studied as these methods require extra handling and may increase the likelihood of duodenoscope structural
RI PT
damage. A limitation of our study was that we did not characterize positive cultures into pathogenic and non-pathogenic bacteria. For reasons mentioned above, we submit that proportion of duodenoscopes showing significant aerobic growth (≥10 CFU) better
SC
reflects the entire disinfection process and is therefore likely to be a more sensitive marker of duodenoscope contamination than proportion of duodenoscopes showing
M AN U
pathogenic bacterial growth. Regardless, is not likely that these two end-points would show discordant results. Thus the absence of information about duodenoscope contamination with pathogenic bacteria is unlikely to have altered the results of our study. It is being increasing recognized that duodenoscope contamination may be
TE D
perpetuated by biofilm that sustains viable bacteria. Biofilm has been identified within endoscope channels following appropriate reprocessing of duodenoscopes (6,23,24 ). Furthermore, studies using adenosine triphosphate to detect organic matter have
EP
suggested that biofilm may account for insensitivity of culture to detect viable bacteria particularly involving the elevator mechanism (25,26). Future studies may therefore
AC C
need to broaden the outcomes measures to include not only culture results, but some measure of biofilms as well.
Several limitations of our study warrant further discussion. First, the Centers for Disease Control recommend that ideally 2 persons perform duodenoscope cultures. In our study duodenoscope cultures were performed by one study personnel, and that may
21
ACCEPTED MANUSCRIPT
be increased the possibility of environmental contamination of the study cultures. Second, there is no consensus on what parts of the sHLD process should be repeated for dHLD. A statement by the FDA suggested that the dHLD process could be
RI PT
conducted “…either manually or through the use of Automated Endoscope
Reprocessors” (27). We defined dHLD as single manual wipe and single manual
reprocessing cycle followed by 2 re-processing cycles of the automated endoscope
SC
reprocessor. An alternative approach to dHLD could have included repeating the
manual wipe and the manual reprocessing cycle twice as well. It is uncertain if the
M AN U
addition of a second cycle of manual reprocessing may have improved the effectiveness of dHLD. Third, we did not characterize positive cultures into pathogenic and nonpathogenic bacteria and are therefore unable to report on the differences between the
TE D
three study arms in this regard.
In summary, we found that in the non-outbreak setting, duodenoscope contamination by MRDOs is extremely uncommon, and this limited the statistical power of our study with
EP
regard to our primary end-point. However, the dHLD and HLD/ETO arms fared worse than the sHLD arm with regard to proportion of duodenoscopes with >10 CFU microbial
AC C
growth (secondary end-point). Our results do not support the routine use of dHLD or ethylene oxide sterilization for duodenoscope reprocessing.
22
ACCEPTED MANUSCRIPT
REFERENCES 1. Carbonne A, Thiolet JM, Fournier S, et al. Control of a multi-hospital outbreak of KPC-producing Klebsiella pneumoniae type 2 in France, September to October
2.
RI PT
2009. Euro Surveill 2010;15. Aumeran C, Poincloux L, Souweine B, et al. Multidrug-resistant Klebsiella
pneumoniae outbreak after endoscopic retrograde cholangiopancreatography.
3.
SC
Endoscopy 2010;42:895-9.
Epstein L, Hunter JC, Arwady MA, et al. New Delhi metallo-beta-lactamase-
M AN U
producing carbapenem-resistant Escherichia coli associated with exposure to duodenoscopes. Jama 2014;312:1447-55. 4.
Wendorf KA, Kay M, Baliga C, et al. Endoscopic Retrograde Cholangiopancreatography-Associated AmpC Escherichia coli Outbreak. Infect Control Hosp Epidemiol 2015;Epub ahead of print (3/30/2015):1-9. Alrabaa S. Early identification and control of carbapenemase-producing
TE D
5.
Klebsiella pneumoniae, originating from contaminated endoscopic equipment.
6.
EP
Am J Infect Control 2013;41:850.
Verfaillie CJ, Bruno MJ, A FVItH, et al. Withdrawal of a novel-design
AC C
duodenoscope ends outbreak of a VIM-2-producing Pseudomonas aeruginosa. Endoscopy 2015;Epub ahead of print (3/31/2015).
7.
U.S. Food and Drug Administration. Design of Endoscopic Retrograde Cholangiopancreatography (ERCP) Duodenoscopes May Impede Effective Cleaning: FDA Safety Communication, 2015.
8.
Petersen BT, Koch J, Ginsberg GG. Infection Using ERCP Endoscopes. Gastroenterology 2016;151:46-50. 23
ACCEPTED MANUSCRIPT
9.
Smith ZL, Oh YS, Saeian K, et al. Transmission of carbapenem-resistant Enterobacteriaceae during ERCP: time to revisit the current reprocessing guidelines. Gastrointest Endosc 2015;81:1041-5. Rutala WA, Weber DJ. Outbreaks of carbapenem-resistant Enterobacteriaceae
RI PT
10.
infections associated with duodenoscopes: What can we do to prevent infections? Am J Infect Control 2016;44:e47-51.
Centers for Disease Control and Prevention. Interim Duodenoscope Culture
SC
11.
Method, 2015.
Chen A LB, Wright SB, Yassa DS, Snyder GM, Hirsch EB. Delayed processing of
M AN U
12.
simulated duodenoscope and patient cultures yields viable bacteria. Society for Healthcare Epidemiology of America Spring 2016 Conference. Atlanta, Georgia, U.S.A., 2016.
Centers for Disease Control and Prevention. Interim Duodenoscope Surveillance
TE D
13.
Protocol, 2015 Available at: http://www.cdc.gov/hai/organisms/cre/creduodenoscope-surveillance-protocol.html. Accessed 1/5/2017 . Ross AS, Baliga C, Verma P, et al. A quarantine process for the resolution of
EP
14.
duodenoscope-associated transmission of multidrug-resistant Escherichia coli.
15.
AC C
Gastrointest Endosc 2015;82:477-83.
Paula H, Presterl E, Tribl B, et al. Microbiologic surveillance of duodenoscope reprocessing at the Vienna University Hospital from November 2004 through March 2015. Infect Control Hosp Epidemiol 2015;36:1233-5.
24
ACCEPTED MANUSCRIPT
16.
Brandabur JJ, Leggett JE, Wang L, et al. Surveillance of guideline practices for duodenoscope and linear echoendoscope reprocessing in a large healthcare system. Gastrointest Endosc 2016;84:392-399 e3. Naryzhny I, Silas D, Chi K. Impact of ethylene oxide gas sterilization of
RI PT
17.
duodenoscopes after a carbapenem-resistant Enterobacteriaceae outbreak. Gastrointest Endosc 2016;84:259-62.
Almario CV, May FP, Shaheen NJ, et al. Cost Utility of Competing Strategies to
SC
18.
Prevent Endoscopic Transmission of Carbapenem-Resistant
19.
M AN U
Enterobacteriaceae. Am J Gastroenterol 2015;110:1666-74.
Chen A, Lane BV, Wright SB, Yassa DS, Snyder GM, Hirsch EB. Delayed processing of simulated duodenoscope and patient cultures yields viable bacteria. Poster presentation, Society for Healthcare Epidemiology of America
20.
TE D
Spring 2016 Conference, Atlanta, GA May 18-21, 2016).
https://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/M
EP
edicalDevices/MedicalDevicesAdvisoryCommittee/GastroenterologyUrologyDevicesPanel/ucm445590.htm. accessed 5/27/17. Rutala WA, Weber DJ. Outbreaks of carbapenem-resistant Enterobacteriaceae
AC C
21.
infections associated with duodenoscopes: What can we do to prevent infections? Am J Infect Control 2016; 44(5 Suppl): e47-51
22.
Heroux R, Sheppard M, Wright SB, Sawhney M, Hirsch EB, Kalaidjian R, Snyder GM. Duodenoscope hang time does not correlate with risk of bacterial contamination. Am J Infect Control. 2016 Dec 26. PMID: 28034537
25
ACCEPTED MANUSCRIPT
23.
Vickery K, Ngo QD, Zou J, Cossart YE. The effect of multiple cycles of contamination, detergent washing, and disinfection on the development of biofilm in endoscope tubing. Am J Infect Control. 2009;37(6):470-5. Pajkos A, Vickery
RI PT
K, Cossart Y. Is biofilm accumulation on endoscope tubing a contributor to the failure of cleaning and decontamination? J Hosp Infect. 2004;58(3):224-9 24.
Pajkos A, Vickery K, Cossart Y. Is biofilm accumulation on endoscope tubing a
SC
contributor to the failure of cleaning and decontamination? J Hosp Infect. 2004;58(3):224-9.
Olafsdottir LB, Wright SB, Smithey A, Herous R, Hirsch EB, Chen A, Lane B,
M AN U
25.
Sawhney MS, Snyder GM. Adenosine Triphosphate Quantification Correlates Poorly with Microbial Contamination of Duodenoscopes. Infect Control Hosp Epidemiol, 2017;17:1-7
Sethi S, Huang RJ, Barakat MT, Banaei N, Friedland S, Banerjee S. Adenosine
TE D
26.
triphosphate bioluminescence for bacteriologic surveillance and reprocessing strategies for minimizing risk of infection transmission by duodenoscopes.
27.
EP
Gastrointest Endosc. 2017 Jun;85(6):1180-1187. https://www.fda.gov/MedicalDevices/Safety/AlertsandNotices/ucm454766.htm.
28.
AC C
Accessed 6/22/17.
Rutala WA, Weber DJ. ERCP scopes: what can we do to prevent infections? Infect Control Hosp Epidemiol. 2015 Jun;36(6):643-8.
26
ACCEPTED MANUSCRIPT
FIGURE 1: Frequency of no growth (0 CFU), low quantity of growth (1-9 CFU) and significant quantity of growth (≥10 CFU) in each of the trial arms Note: CFU, colony-forming units; ETO, ethylene oxide gas sterilization; HLD, high-level
RI PT
disinfection
FIGURE 2: Culture results from individual study duodenoscopes for each day
SC
sampled during the study period.
Note: CFU, colony-forming units; dHLD, double high-level disinfection; HLD/ETO, high-
M AN U
level disinfection/ethylene oxide gas sterilization; sHLD, standard high-level disinfection. The x-axis represents all dates during the study period for which one or more duodenoscopes were cultured; the y-axis represents the 18 (de-identified) duodenoscopes. Empty cells indicate no culture was taken for that duodenoscope on
TE D
that day. The number in each cell represents the highest CFU of aerobic bacterial growth from either the elevator mechanism or working channel, including in cases
AC C
EP
where a duodenoscope was cultured more than once on a given day.
27
ACCEPTED MANUSCRIPT
TABLE 1: Frequency of the primary outcome (≥1 multidrug-resistant organism), or secondary outcomes of any growth > 0 CFU and growth of ≥10 CFU on any duodenoscope culture
RI PT
Growth, elevator mechanism or working channel (%) (N)
≥1 MDRO
>0 CFU*
≥10 CFU†
sHLD
174
0
28 (16.1)
4 (2.3)
dHLD
169
0
27 (16.0)
7 (4.1)
HLD/ETO 173
0
39 (22.5)
0
94 (18.3)
516
9 (4.2)
20 (3.9)
M AN U
Total
SC
Trial Arm
Note: CFU, colony-forming units; dHLD, double high-level disinfection; HLD/ETO, standard high-level disinfection followed by ethylene oxide gas sterilization; MDRO, multidrug-resistant organism; sHLD, standard high-level disinfection.
TE D
See Methods and the Supplementary Material for definition of MDRO.
AC C
EP
* p = 0.21, † p = 0.36 by Fisher’s exact test
ACCEPTED MANUSCRIPT
TABLE 2: Association between trial arm and duodenoscope contamination in generalized estimating equations regression models Duodenoscope
contamination with >0 CFU
contamination with ≥10 CFU
Variable
RR (95%CI)
sHLD
ref
p-value
Unadjusted
RR (95%CI)
p-value
ref
SC
—
model
RI PT
Duodenoscope
—
1.02 (0.66-1.58)
0.92
2.05 (1.06-3.94)
0.03
HLD/ETO
1.43 (0.89-2.29)
0.14
2.40 (1.13-5.09)
0.02
Adjusted sHLD
ref
model*
M AN U
dHLD
—
—
—
1.02 (0.66-1.58)
0.93
— —
—
HLD/ETO
1.41 (0.86-2.32)
0.17
— —
—
TE D
dHLD
Note: 95%CI, 95% confidence interval; CFU, colony-forming units; sHLD, standard highlevel disinfection; dHLD, double high-level disinfection; HLD/ETO, high-level disinfection
EP
followed by ethylene oxide gas sterilization; RR, relative risk. The adjusted model included duration of preceding procedure (in hours), use of sphincterotomy and stent
AC C
removal during the preceding procedure, and time elapsed between preceding procedure and repeat culture (in days), all non-significant (full regression models are presented in the Supplementary Material).
ACCEPTED MANUSCRIPT
TABLE 3: Characteristics of patients enrolled in study. Patients undergoing procedures with a duodenoscope assigned to a study arm
N, (%)
All patients
RI PT
Characteristic
HLD/ETO (N=158)
dHLD
sHLD
(N=166)
(N=171)
90 (54.2)
81 (47.4)
66.5 (16.3)
63.0 (17.5)
62.8 (15.9)
27.7 (6.9)
27.9 (6.5)
27.3 (5.5)
19 (12.0)
26 (15.7)
14 (8.2)
(N=495)*
SC
Patient characteristics 236 (47.7)
Age, mean, years (SD)
64.0 (16.6)
Body mass index, mean, kg/m2 (SD)
27.6 (6.3)
Hepatic or pancreaticobiliary mass
59 (11.9)
Biliary abnormality†
181 (36.6)
60 (38.0)
66 (39.8)
55 (32.2)
Cholangitis
46 (9.3)
13 (8.2)
9 (5.4)
24 (14.0)
Solid organ transplant‡
16 (3.2)
6 (3.8)
6 (3.6)
4 (2.3)
Hematologic malignancy
6 (1.2)
0
1 (0.6)
5 (2.9)
124 (25.1)
41 (26.0)
39 (23.5)
44 (25.7)
5 (1.0)
2 (1.3)
1 (0.6)
2 (1.2)
Chronic hemodialysis Procedure characteristics
TE D
EP
AC C
Diabetes mellitus, type 1 or 2
65 (41.1)
M AN U
Female gender
ACCEPTED MANUSCRIPT
Procedure duration, mean, minutes
32.9 (18.9)
33.7 (20.8)
Sphincterotomy
208 (42.0)
67 (42.4)
Stent placement
187 (37.8)
61 (38.6)
Stent removal
132 (26.7)
41 (25.6)
Balloon dilation or stone extraction
301 (60.8)
99 (62.7)
Brushing for cytology
71 (14.3)
32.4 (18.4)
32.7 (17.6)
M AN U
SC
RI PT
(SD)
18 (11.4)
Full demographic and procedure data unavailable for 3 of 498 patients
†
Biliary duct obstruction, stenosis, or leak
‡
Liver (14), kidney (3); one patient had both a liver and kidney transplant
TE D
*
65 (39.2)
76 (44.4)
60 (36.1)
66 (38.6)
46 (27.7)
45 (26.3)
104 (62.7)
98 (57.3)
24 (14.5)
29 (17.0)
Note: Patients undergoing multiple procedures are included for each procedure that was performed during the study
AC C
EP
period. See text for details regarding frequency of repeat procedures.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S0016-5085(17)35869-9 10.1053/j.gastro.2017.06.052 YGAST 61271
To appear in: Gastroenterology Accepted Date: 27 June 2017 Please cite this article as: Snyder GM, Wright SB, Smithey A, Mizrahi M, Sheppard M, Hirsch EB, Chuttani R, Heroux R, Yassa DS, Olafsdottir LB, Davis RB, Jiannis A, Bapat V, Bidari K, Pleskow DK, Leffler D, Lane B, Chen A, Gold HS, Bartley A, King AD, Sawhney MS, Randomized Comparison of 3 High-level Disinfection and Sterilization Procedures for Duodenoscopes, Gastroenterology (2017), doi: 10.1053/j.gastro.2017.06.052. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
The DISINFECTS trial is a prospective randomized trial investigating three methods of reprocessing duodenoscopes: standard high-level disinfection with ortho-phthalaldehyde disinfectant, standard highlevel disinfection with a repeated (double) cycle of disinfectant exposure, and standard high-level disinfection followed by ethylene oxide gas sterilization. No significant difference was noted between these reprocessing methods with regard to proportion of duodenoscopes with MDRO contamination, or demonstrating ≥10 CFU or >0 CFU growth of any aerobic bacteria.
ACCEPTED MANUSCRIPT
Title: Randomized Comparison of 3 High-level Disinfection and Sterilization
Short Title: Duodenoscope Disinfection Trial
RI PT
Procedures for Duodenoscopes
Authors: Graham M. Snyder MD SM1,2, Sharon B. Wright MD MPH 1,2, Anne
SC
Smithey1,4, Meir Mizrahi MD3, Michelle Sheppard RN MSN DNPc3, Elizabeth B. Hirsch PharmD4, Ram Chuttani MD2,3, Riley Heroux1,4, David S. Yassa MD MPH1,2, Lovisa B.
M AN U
Olafsdottir MD1, Roger B. Davis ScD5, Anastasiou Jiannis MD3, Vijay Bapat MD3, Kiran Bidari MD3, Douglas K. Pleskow MD2,3, Daniel Leffler MD MS2,3, Benjamin Lane BS1, Alice Chen1,4, Howard S Gold MD1,2, Anthony Bartley MD3, Aleah D. King RN BSN1, Mandeep S. Sawhney MBBS MS2,3
TE D
1. Division of Infection Control/Hospital Epidemiology, Beth Israel Deaconess Medical Center, Boston, Massachusetts
2. Harvard Medical School, Boston, Massachusetts
Massachusetts
EP
3. Division of Gastroenterology, Beth Israel Deaconess Medical Center, Boston,
AC C
4. Department of Pharmacy and Health Systems Sciences, Northeastern University, Boston, Massachusetts
5. Division of General Medicine and Primary Care, Beth Israel Deaconess Medical Center, Boston, Massachusetts
Grant support:
1
ACCEPTED MANUSCRIPT
This work was funded in part by an investigator-initiated grant from the American Society for Gastrointestinal Endoscopy (G.M.S., M.S.S.) and from Beth Israel Deaconess Medical Center, Boston, Massachusetts. Study sponsors did not play any
RI PT
role in the study design, collection, analysis, and interpretation of data.
Abbreviations
SC
CFU, colony-forming units
HLD, high-level disinfection
M AN U
dHLD, double high-level disinfection
HLD/ETO, standard high-level disinfection followed by ethylene oxide gas sterilization MDRGN, multidrug-resistant Gram-negative pathogens MDRO, multidrug resistant organisms
TE D
MRSA, methicillin-resistant Staphylococcus aureus sHLD, standard high-level disinfection
EP
VRE, vancomycin-resistant enterococci
Main Outcomes and Measures: The primary outcome was the proportion of
AC C
duodenoscopes with an elevator mechanism or working channel culture showing ≥1 MDRO; secondary outcomes included the frequency of duodenoscope contamination with >0 and ≥10 colony-forming units (CFU)
Corresponding Author: Mandeep Sawhney, MBBS MS
2
ACCEPTED MANUSCRIPT
330 Brookline Ave, Rabb-Rose 133 Boston, MA 02215, USA Phone: (617) 667-1088
RI PT
Fax: (617) 667-1171 Email: [email protected]
SC
Disclosures:
M AN U
Conflicts of Interest. All authors report no other potential conflicts of interest to disclose.
Author Contributions:
Conception/design: GMS, SBW, AS, MM, EBH, DSY, RBD, DL, HSG, AB, ADK, MS Data collection: AS, MM, MS, RC, RH, LBO, AJ, VB, KB, DKP, AB, MS
TE D
Data analysis: GMS, AS, EBH, RH, LBO, RBD, BL, AC Manuscript preparation: GMS, SBW, MSS, RBD
EP
Approval of final version of the manuscript: all authors
ACKNOWLEDGEMENTS
AC C
Portions of this data have been presented at the Society for Healthcare Epidemiology of America Spring 2016 Conference and at Digestive Disease Week 2017. This work was conducted with support from Harvard Catalyst (NIH Award UL1 TR001102).
The authors would like to extend our appreciation to James Kirby, M.D. for his advice on the design of the study, Nancy Doraiswami and Maxine Hainstock for facilitating
3
ACCEPTED MANUSCRIPT
additional reprocessing required for the study, and Christopher Rowley, M.D. and Alan Moss, M.D. for their advising as independent members of the Safety Review
AC C
EP
TE D
M AN U
SC
Endoscopy Unit for their work and support on the project.
RI PT
Committee. Additionally, we would like to thank the staff of the Stoneman 4 Advanced
4
ACCEPTED MANUSCRIPT
ABSTRACT Background and Aims: Duodenoscopes have been implicated in the transmission of multi-drug resistant bacteria (MDRO). We compared the frequency of duodenoscope
RI PT
contamination with MDRO or any other bacteria after disinfection or sterilization by 3 different methods.
SC
Methods: We performed a single-center prospective randomized study in which
duodenoscopes were randomly reprocessed by standard high-level disinfection (sHLD),
M AN U
double high-level disinfection (dHLD), or standard high-level disinfection followed by ethylene oxide gas sterilization (HLD/ETO). Samples were collected from the elevator mechanism and working channel of each duodenoscope and cultured before use. The primary outcome was the proportion of duodenoscopes with an elevator mechanism or
TE D
working channel culture showing 1 or more MDRO; secondary outcomes included the frequency of duodenoscope contamination with more than 0 and 10 or more colony-
EP
forming units (CFU) of aerobic bacterial growth on either sampling location.
Results: After 3 months of enrollment, the study was closed due to the futility—we did
AC C
not observe sufficient events to evaluate the primary outcome. Among 541 duodenoscope culture events, 516 were included in the final analysis. No duodenoscope culture in any group was positive for MDRO. Bacterial growth of more than 0 CFU was noted in 16.1% duodenoscopes in the sHLD group, 16.0% in the dHLD group, and 22.5% in the HLD/ETO group (P=.21). Bacterial growth or 10 or more CFU was noted in 2.3% of duodenoscopes in sHLD group, 4.1% in the dHLD group, and
5
ACCEPTED MANUSCRIPT
4.2% in the HLD/ETO group (P=.36). MRDOs were cultured from 3.2% of pre-procedure rectal swabs and 2.5% of duodenal aspirates.
RI PT
Conclusions: In a comparison of duodenoscopes reprocessed by sHLD, dHLD, or
HLD/ETO, we found no significant differences between groups for MDRO or bacteria contamination. Enhanced disinfection methods (dHLD or HLD/ETO) did not provide
SC
additional protection against contamination. However, insufficient events occurred to
M AN U
assess our primary study end-point. ClinicalTrials.gov no: NCT02611648
AC C
EP
TE D
KEY WORDS: DISINFECTS study; surveillance; endoscopy; non-outbreak setting
6
ACCEPTED MANUSCRIPT
INTRODUCTION Recent outbreaks of infections due to carbapenem-resistant Enterobacteriaceae and other multidrug-resistant organisms (MDRO) have been attributed to contaminated
RI PT
duodenoscopes, despite reported compliance with manufacturer recommendations for reprocessing.1-6 The reasons for duodenoscope contamination when reprocessing is conducted appropriately are not fully understood, and may relate to biofilm formation
SC
within the duodenoscope.
An advisory panel convened by the Food and Drug Administration agreed that
M AN U
current standards for duodenoscope reprocessing did not provide a reasonable level of safety and effectiveness.7 Experts have suggested the adoption of additional measures like double reprocessing cycles and ethylene oxide sterilization to supplement current standards.8-10 These additional measures are expensive, time consuming and not
TE D
readily available in all institutions. Furthermore, the effectiveness of these additional measures has never been systematically studied in a non-outbreak setting. Studies comparing different methods of duodenoscope reprocessing are therefore needed if we
EP
are to identify the optimal method of duodenoscope reprocessing. In the DISINFECTS trial (Duodenoscope Infection Surveillance IN Functioning
AC C
automated Endoscope reprocessors in Conjunction with eThlyene oxide Sterilization), we aimed to compare the frequency of duodenoscope contamination with MDRO and with any bacterial contamination after reprocessing using three methods of high-level disinfection or sterilization.
METHODS
7
ACCEPTED MANUSCRIPT
Study design and setting We conducted a prospective randomized trial investigating three methods of reprocessing duodenoscopes: standard high-level disinfection with ortho-
RI PT
phthalaldehyde disinfectant (sHLD), standard high-level disinfection with a repeated (double) cycle of disinfectant exposure (dHLD), and standard high-level disinfection followed by ethylene oxide gas sterilization (HLD/ETO). The study was approved by the
SC
Institutional Review Board at our institution, and was registered at clinicaltrials.gov
(NCT02611648). All authors had access to study data and reviewed and approved the
M AN U
final manuscript.
During the study period, all duodenoscopes selected for patient use were cultured prior to the procedure. Patients undergoing an endoscopy procedure for which the use of a duodenoscope was anticipated were offered participation in the study. Prior
TE D
to patient enrollment, 18 duodenoscopes were assigned to one of three study arms. Reprocessing time for duodenoscopes assigned to the HLD/ETO arm was estimated to be substantially longer than the other arms; we therefore assigned eight
EP
duodenoscopes to the HLD/ETO arm, and five each to the sHLD and dHLD study arm. Duodenoscope assignment to a study arm was retained for the duration of the study. A
AC C
permuted block randomization key specifying study arm was randomly generated with a block size of three (comprising one from each study arm). Endoscopy procedures were performed according to standard clinical practice. The attending endoscopist and patient were blinded to duodenoscope study arm assignment.
8
ACCEPTED MANUSCRIPT
Patient history of MDRO carriage was obtained from Infection Control databases at the study institution. At the time of the procedure, a rectal swab and duodenal aspirate sample were obtained to identify carriage with MDRO.
RI PT
Study Procedures Devices and reprocessing
Duodenoscope reprocessing during the study entailed a manual wipe of the
SC
exterior of the duodenoscope with enzymatic solution (EmPower, Metrex) immediately following the procedure, manual reprocessing within one hour of the procedure
M AN U
including the use of device-specific elevator mechanism and working channel brushes, and completion of reprocessing using automated endoscope reprocessors (System 83 Plus 9, Custom Ultrasonics) with ortho-phthalaldehyde disinfectant (MetriCide OPA Plus, Metrex) followed by ethanol flush and vertical hanging for drying in a non-
TE D
ventilated cabinet. Duodenoscopes assigned to the dHLD arm, underwent reprocessing similar to the sHLD arm, except 2 re-processing cycles of automated endoscope reprocessor with ortho-phthalaldehyde disinfectant were used. For duodenoscopes
EP
assigned to the HLD/ETO trial arm, ethylene oxide gas sterilization was performed following these reprocessing steps (Steri-Vac Sterilizer/Aerator, 3M). Prior to study
AC C
initiation, our endoscopy unit utilized seven duodenoscopes (Olympus, model TJFQ180V) all of which were acquired in 2012; in order to accommodate additional reprocessing time associated with our study, an additional 11 new duodenoscopes of the same manufacturer and model were acquired in 2015. The 2012 and 2015 duodenoscopes were proportionally distributed to study arms. Duodenoscope and patient sample collection and microbiological methods
9
ACCEPTED MANUSCRIPT
The duodenoscope culturing process consisted of two samples obtained using stringent aseptic technique: 1) a sterile dry flocked swab sample of the elevator mechanism, and 2) a sterile saline sample of the working channel using the “flush-
RI PT
brush-flush” technique.11 The flocked swab elevator mechanism sample was placed into transport media (ESwab with liquid Amies media, Copan Diagnostics). The working channel sample was obtained by flushing sterile water through the working channel,
SC
passing a sterile device-specific brush (Olympus) through the length of the working channel and manually agitated in the captured sterile water after removal, and then
M AN U
flushing another aliquot of sterile water through the working channel and captured with the initial flush specimen. A standardized volume of sample was added to phosphatebuffered saline for the transport of working channel samples.12 Duodenoscopes were sampled in a dedicated location. Duodenoscopes cultured but not used within one
TE D
calendar day following culture were reprocessed without patient use. Duodenoscope sample cultures were processed for two outcomes: growth of any aerobic bacteria (colony-forming units, CFU), and the presence of MDRO. MDRO were defined as
EP
methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), or multidrug-resistant Gram-negative pathogens (MDRGN).
AC C
Among patients providing informed consent for the patient microbiologic
component of the study, a rectal swab (CultureSwab Plus Amies Gel without Charcoal, BD) and duodenal aspirate were obtained at the time of the procedure. Patients were given the opportunity to opt-out of the rectal swab collection. Patient samples were processed for the presence of MDRO. The detailed duodenoscope and patient sample collection and microbiologic methods can be found in the Supplementary Material.
10
ACCEPTED MANUSCRIPT
Data collection and safety monitoring Patient- and procedure-related information was collected by study investigators using a structured intake form developed for the study. An independent observer
RI PT
blinded to trial arm assignment reviewed duodenoscope culture results each weekday to identify duodenoscopes that met pre-specified criteria for immediate review by the Safety Monitoring Committee. Blinded data were reviewed in relation to potentially
SC
defective duodenoscopes and loss of intervention arm equipoise. Statistical analysis
M AN U
The final analysis included all procedures in which a randomly assigned duodenoscope was used on a patient and both the elevator mechanism and working channel were sampled. The primary outcome of the study is the proportion of duodenoscopes in each of the three disinfection arms contaminated with ≥1 MDRO on
TE D
either the elevator mechanism or working channel. Secondary outcomes included the proportion contaminated with ≥10 CFU and >0 CFU of any aerobic bacteria on either the elevator mechanism or working channel in the three disinfection arms, and the
EP
frequency of patient MDRO acquisition after exposure to a duodenoscope with MDRO contamination. Differences in proportions of events among study arms were calculated
AC C
using Fisher’s exact test. We used a group sequential design with a Haybittle stopping rule to allow monthly interim analyses while preserving a study-wide 5% level of significance for the primary outcome. Blinded results were presented to the Safety Monitoring Committee, which made recommendations concerning continuation of the study.
11
ACCEPTED MANUSCRIPT
To explore factors associated with duodenoscope contamination with either ≥1 MDRO, any growth (>0 CFU) or significant contamination (≥10 CFU), on either the elevator mechanism or working channel, we fit Poisson regression models with trial arm
RI PT
as the factor of interest to obtain estimates of relative risk between standard disinfection and each of the more intensive methods. Models were fit using generalized estimating equations methods with an autoregressive correlation structure to account for
SC
correlation among observations from the same duodenoscope. To address for
confounding by factors other than trial arm that may be associated with duodenoscope
M AN U
contamination, covariates included in an adjusted model included duration of preceding procedure (in hours), use of sphincterotomy and stent removal during the preceding procedure, and time elapsed between preceding procedure and repeat culture (in days). We calculated a sample size assuming 5% contamination in the standard
TE D
practice trial arm (sHLD) and 0.5% contamination in the intensive intervention trial arm (HLD/ETO). We set the level of significance to 5% and power to 70%. With adjustments for the 3-arm design (using the least favorable distribution assumption) and the group
AC C
RESULTS
EP
sequential design, a sample size of 206 per trial arm was required.
After two months of study enrollment, the Safety Review Committee
recommended that patient microbiologic data collection be stopped due to very low probability of observing patient MDRO acquisition events, and after the third month of enrollment recommended that the study close due to the futility of observing sufficient events to evaluate the primary outcome.
12
ACCEPTED MANUSCRIPT
Duodenoscope contamination Among the 541 study observations, 531 (98.2%) duodenoscopes were cultured prior to the procedure, 517 (95.6%) duodenoscopes were used following appropriate
RI PT
randomization, and 516 (95.4%) were cultured and appropriately randomized prior to selection for a patient procedure. These 516 observations intended for patient use were included in the primary analysis including 174 (33.7%) assigned to the sHLD arm, 169
SC
(32.8%) to the dHLD arm, and 173 (33.5%) to the HLD/ETO arm.
No duodenoscope cultures grew ≥1 MDRO from either the elevator mechanism
M AN U
or working channel. Thus there was no difference between the three study arms with regard to the primary study outcome (Table 1). Twenty (3.9%) duodenoscope cultures demonstrated growth of ≥10 CFU in either the elevator mechanism or working channel, including 11 (2.0%) on the elevator mechanism only, eight (1.5%) on the working
TE D
channel, and one (0.2%) on both the elevator mechanism and working channel. There was no difference between the three study arms with regard to >0 CFU or ≥10 CFU growth of any aerobic bacteria (Table 1). Figure 1 demonstrates the frequency
EP
of no growth (0 CFU), low quantity of growth (1-9 CFU) and significant quantity of growth (≥10 CFU) in each of the study arms. Figure 2 demonstrates the pattern of
AC C
positive cultures on individual duodenoscopes throughout the study period. No duodenoscopes met pre-specified criteria for persistent duodenoscope contamination. Duodenoscope contamination regression modeling A regression model demonstrated no association between study arm and
frequency of duodenoscope contamination when >0 CFU bacterial growth was used as the outcome variable (Table 2). No significant association between the enhanced study
13
ACCEPTED MANUSCRIPT
arms and contamination was noted even after adjusting for duration of preceding procedure, sphincterotomy, stent removal, and time elapsed between preceding procedure and repeat culture. The regression model with the outcome of ≥10 CFU
RI PT
growth modeled without covariates demonstrated an approximately 2-fold higher
likelihood of duodenoscope contamination among the dHLD and HLD/ETO arms compared to the sHLD arm (Table 2).
SC
Duodenoscope exposure to patient carriage of MDRO
For 43 of the 541 (7.9%) study observations the duodenoscope did not have
M AN U
patient contact as either >1 calendar day had passed between the time of culture and demand for duodenoscope use (n=16), or duodenoscopes were obtained for use but did not subsequently have patient contact (n=27). The remaining 498 procedures were performed on 389 patients, including three (0.8%) patients who underwent four
TE D
procedures, 15 (3.9%) patients who underwent three procedures, and 70 (18.0%) patients who underwent two procedures. Among the 389 patients undergoing at least one procedure during the study, 26 patients had a known history of carriage with ≥1
EP
MDRO prior to study participation. This included MRSA, 15 (3.9%); VRE, 12 (3.1%); and MDRGN, 9 (2.3%). Three patients had a history of both MRSA and VRE, one
AC C
patient with both VRE and MDRGN, and three patients with a history of MRSA, VRE, and MDRGN carriage.
For 343 of the 498 (68.9%) procedures patients were offered participation in the
patient microbiologic component of the study. Among these 343 procedures, 260 (75.8%) patients provided full consent (236 of 260 patients) or opted out of collection of rectal samples (24 of 260 patients). Twenty-four (7.0%) patients declined to participate,
14
ACCEPTED MANUSCRIPT
and for 59 (17.2%) patients consent was not sought due to procedure urgency. The remaining 155 of 498 (31.1%) patients underwent procedures after discontinuation of the patient microbiologic component of the study. The demographic characteristics of all
RI PT
study patients including patients undergoing procedures with duodenoscopes from each study arm are presented in Table 3.
Among the 260 consenting patients, rectal sampling for MDRO was obtained
SC
among 189 (72.7%) and duodenal aspirates among 202 (77.7%). On rectal swabs,
MRSA, VRE, and MDRGN were identified among one (0.5%), three (1.6%), and three
M AN U
(1.6%) of samples, respectively; a total of six samples (3.2%) had ≥1 MDRO including one patient with co-carriage of VRE and MDRGN. Among duodenal aspirates, MRSA, VRE, and MDRGN were identified among two (1.0%), two (1.0%), and one (0.5%) of
DISCUSSION
TE D
samples, respectively, including one MDRO for each of these five (2.5%) samples.
We report the results of the first prospective randomized trial comparing sHLD, dHLD
EP
and HLD/ETO for reprocessing duodenoscopes. In a non-outbreak setting we found no significant difference between these methods with regard to proportion of
AC C
duodenoscopes with MDRO contamination, or demonstrating ≥10 CFU or >0 CFU growth of any aerobic bacteria. In our study, MDRO were not recovered from any duodenoscope culture, and this limited our ability to detect differences between study arms with regard to the primary study end-point.
15
ACCEPTED MANUSCRIPT
The clinical events that prompted this trial were the reported outbreaks of patients acquiring MRDOs from contaminated duodenoscopes (1-6). We are unaware of any recent published reports of non-MDRO related duodenoscope acquired infections. We
RI PT
therefore focused on MDROs as a clinically relevant primary study end-point. When we designed the study in 2015 there were no published data on the proportion of
duodenoscopes contaminated with MRDOs. Using a conservative approach based on
SC
studies characterizing high-pathogenicity organisms, we powered our study for 5% of duodenoscopes in the sHLD arm showing contamination with MRDOs. Our study, in a
M AN U
tertiary care referral center, found that with vigilance in reprocessing MDRO contamination did not occur among any 516 duodenoscope culture events. Similar results have since been reported by Brandabur et al, who found 0 of 4032 duodenoscope cultures positive for MRDOs (16). In our study, 3.2% of rectal swabs
TE D
and 2.5 % of duodenal aspirates showed colonization by MRDO. These findings suggest that in the non-outbreak setting, sHLD may be able to adequately disinfect duodenoscopes even when endoscopes are used for procedures on a proportion of
EP
patients with MDRO colonization.
AC C
Duodenoscope contamination, regardless of the pathogenicity of the organism or moment of contamination, reflects a possible defect in reprocessing, handling and storage, or the device itself that may place the patient at risk of transmitted MDRO. To account for possible contamination from the time the duodenoscope was withdrawn from a patient until it was re-inserted into the next patient, we determined that significant bacterial contamination immediately prior to use would be more representative of this
16
ACCEPTED MANUSCRIPT
entire process than simply the presence or absence of pathogenic bacteria. We therefore used growth of ≥10 CFU as recommended by the Centers for Disease Control and Prevention to indicate potentially significant contamination (13). We also analyzed
RI PT
as an outcome any growth (> 0 CFU) from duodenoscope samples. Duodenoscope sampling was conducted in a dedicated space, and was performed by trained study personnel based using a technique based upon CDC sampling guidelines (13). Culture
SC
and lab techniques were validated using a standardized inoculum in the usual fashion. Additionally, we have previously demonstrated that our method of specimen storage
M AN U
and transport is unlikely to affect our findings (19). Following sHLD, the proportion of duodenoscopes with >0 CFU growth in our study was similar to growth rates reported by others, and thus provides a measure of validity to our culture methods (14-16). Considering only duodenoscopes assigned to the sHLD arm in our study, 16.1% (95%
TE D
CI, 11.2% – 22.1%) showed microbial growth > 0 CFU. Ross et al. reported that 13.1% of cultures collected from duodenoscopes after sHLD were positive for any bacterial growth (14). Brandabur et al reported on curvilinear echoendoscopes and
EP
duodenoscope cultures following sHLD and found up to 14.3% of Olympus endoscopes showed microbial growth (16). The average growth rate for any organism was 8.4% in
AC C
that study. There was very large variation amongst study sites with culture positive rates ranging from 0% to 35%.
At the Gastroenterology and Urology Devices Panel of the Medical Devices Advisory Committee meeting convened by the FDA in 2015, experts strongly urged the FDA to mandate that all duodenoscopes, and preferably all gastrointestinal endoscopes, be
17
ACCEPTED MANUSCRIPT
sterilized by 2018 (20). The FDA has not mandated ethylene oxide sterilization of duodenoscopes, but the use of ethylene oxide sterilization has been strongly proposed (21). The results of our trial do not support the use of ethylene oxide sterilization in a
RI PT
non-outbreak setting. The absence of any duodenooscope cultures showing MDROs lowered the statistical power of our study to detect difference between HLD/ETO and other study arms with regard to our primary end-point. However, the HLD/ETO arm in
SC
fact appeared inferior to sHLD with regard to our secondary end-point (proportion of duodenscope with ≥10 CFU growth). It is highly unlikely that even if we had continued
M AN U
to recruit subjects into the study, this trend would have reversed in the opposite direction to favor the HLD/ETO arm. In the HLD/ETO arm of our study, we found 22.5% of duodenoscopes showed microbial growth >0 CFU, of which 4.2 % showed microbial growth ≥10 CFU; compared with the sHLD arm in which only 16.1% of duodenoscopes
TE D
showed microbial growth >0 CFU, of which 2.3 % showed microbial growth ≥10 CFU. In our regression modeling, duodenoscopes assigned to the HLD/ETO arm in our study were almost twice as likely to show ≥10 CFU microbial growth when compared with the
EP
sHLD arm, and this difference was statistically significant (Table 2). There was also a non-significant increase in the probability of any microbial growth (>0 CFU) in the
AC C
HLD/ETO arm. Similarly, the dHLD arm also appeared have a high proportion of duodenoscopes with > 10 CFU microbial growth when compared with the sHLD arm (4.1% versus 2.3%). Apart from our study, we found only one other study that reported on duodenoscope cultures following ethylene oxide sterilization. Following an outbreak of E. coli at their institution, Naryzhny et al. instituted a policy of HLD/ETO for all duodenoscopes and found that 1.2% remained contaminated with high risk organisms
18
ACCEPTED MANUSCRIPT
(17). This contamination rate is similar to contamination rates seen with sHLD, again suggesting that HLD/ETO may not be superior to sHLD for disinfecting duodenoscopes.
RI PT
Some experts have suggested that ethylene oxide sterilization may increase likelihood of damage to duodenoscopes, and we hypothesize that this may have played a role in the increased duodenoscope contamination noted in the HLD/ETO arm (28). We did not
SC
find an association between the risk of duodenoscope contamination and duration of preceding procedures or performance of stent removals to suggest that instrumentation
M AN U
of the biopsy channel may have played a role in duodenoscope damage and hence contamination. Ross et al found that 3 of 4 duodenoscope that were contaminated with drug-resistant E coli were damaged and needed critical repairs (14). Interestingly, those duodenoscopes did not demonstrate any functional problems and those defects
TE D
were not recognized until the duodenoscopes were returned to the manufacturer. It is not known if dHLD or HLD/ETO sterilization can reliably disinfect structurally defective duodenoscopes. It is also possible that increased contamination seen in the HLD/ETO
EP
arm may be related to the increased handling of the duodenoscopes in that arm. However, the proportion of duodenoscopes with > 10 CFU growth was almost the same
AC C
between the HLD/ETO and the dHLD arm (4.2% versus 4.1%). No extra handling of duodenoscopes is likely to have taken place in the dHLD arm as the dHLD arm was similar to the sHLD arm except that the duodenoscopes underwent an extra automated reprocessing cycle. Differences in duodenoscope storage time between the study groups are also an unlikely explanation for these results. Recently our group has shown that there is no associated between duodenoscope storage time and risk of
19
ACCEPTED MANUSCRIPT
bacterial contamination (22). Ethylene oxide sterilization has several other drawbacks including a substantial increase in cost. HLD using an automated endoscope reprocessor can be performed in less than one hour. Ethylene oxide sterilization
RI PT
increases the duodenoscope processing time to 16 - 24 hours, thus requiring the
purchase of additional duodenoscopes. Naryzhny et al. estimated that using ethylene oxide sterilization increased their institutional expenditure by $ 93,563 over an 18
SC
months period. This estimate did not include additional personnel and increased
duodenoscope maintenance costs (17). In a cost-effective analysis, Almario et al.
M AN U
estimated that the use of ethylene oxide sterilization would increase the cost of each ERCP by $ 1043 (18). Considering that more than 600,000 ERCPs are performed in the United States annually, this would add more than half billion dollars to health care expenditure. Based upon these factors and our study results, we do not support the
TE D
routine use of ethylene oxide sterilization for duodenoscope reprocessing.
Our study has significant implications for further research in this field. With careful
EP
sHLD the proportion of duodenoscopes contaminated with MRDOs, as shown by our study, is likely to be extremely small. Therefore future studies conducted in non-
AC C
outbreak setting should not use MRDOs as the primary outcome, and instead use a surrogate marker of duodenoscope contamination. Proportion of duodenoscopes contaminated with pathogenic bacteria, proportion of duodenoscopes demonstrating significant aerobic growth (≥10 CFU), or proportion of duodenoscope with any aerobic bacteria growth (>0 CFU) are options for such a surrogate end-point. The optimal surrogate end-point should reflect duodenoscope reprocessing, handling and storage,
20
ACCEPTED MANUSCRIPT
and the possible structural damage to the duodenoscope itself. This is especially important when enhanced methods of disinfection are being studied as these methods require extra handling and may increase the likelihood of duodenoscope structural
RI PT
damage. A limitation of our study was that we did not characterize positive cultures into pathogenic and non-pathogenic bacteria. For reasons mentioned above, we submit that proportion of duodenoscopes showing significant aerobic growth (≥10 CFU) better
SC
reflects the entire disinfection process and is therefore likely to be a more sensitive marker of duodenoscope contamination than proportion of duodenoscopes showing
M AN U
pathogenic bacterial growth. Regardless, is not likely that these two end-points would show discordant results. Thus the absence of information about duodenoscope contamination with pathogenic bacteria is unlikely to have altered the results of our study. It is being increasing recognized that duodenoscope contamination may be
TE D
perpetuated by biofilm that sustains viable bacteria. Biofilm has been identified within endoscope channels following appropriate reprocessing of duodenoscopes (6,23,24 ). Furthermore, studies using adenosine triphosphate to detect organic matter have
EP
suggested that biofilm may account for insensitivity of culture to detect viable bacteria particularly involving the elevator mechanism (25,26). Future studies may therefore
AC C
need to broaden the outcomes measures to include not only culture results, but some measure of biofilms as well.
Several limitations of our study warrant further discussion. First, the Centers for Disease Control recommend that ideally 2 persons perform duodenoscope cultures. In our study duodenoscope cultures were performed by one study personnel, and that may
21
ACCEPTED MANUSCRIPT
be increased the possibility of environmental contamination of the study cultures. Second, there is no consensus on what parts of the sHLD process should be repeated for dHLD. A statement by the FDA suggested that the dHLD process could be
RI PT
conducted “…either manually or through the use of Automated Endoscope
Reprocessors” (27). We defined dHLD as single manual wipe and single manual
reprocessing cycle followed by 2 re-processing cycles of the automated endoscope
SC
reprocessor. An alternative approach to dHLD could have included repeating the
manual wipe and the manual reprocessing cycle twice as well. It is uncertain if the
M AN U
addition of a second cycle of manual reprocessing may have improved the effectiveness of dHLD. Third, we did not characterize positive cultures into pathogenic and nonpathogenic bacteria and are therefore unable to report on the differences between the
TE D
three study arms in this regard.
In summary, we found that in the non-outbreak setting, duodenoscope contamination by MRDOs is extremely uncommon, and this limited the statistical power of our study with
EP
regard to our primary end-point. However, the dHLD and HLD/ETO arms fared worse than the sHLD arm with regard to proportion of duodenoscopes with >10 CFU microbial
AC C
growth (secondary end-point). Our results do not support the routine use of dHLD or ethylene oxide sterilization for duodenoscope reprocessing.
22
ACCEPTED MANUSCRIPT
REFERENCES 1. Carbonne A, Thiolet JM, Fournier S, et al. Control of a multi-hospital outbreak of KPC-producing Klebsiella pneumoniae type 2 in France, September to October
2.
RI PT
2009. Euro Surveill 2010;15. Aumeran C, Poincloux L, Souweine B, et al. Multidrug-resistant Klebsiella
pneumoniae outbreak after endoscopic retrograde cholangiopancreatography.
3.
SC
Endoscopy 2010;42:895-9.
Epstein L, Hunter JC, Arwady MA, et al. New Delhi metallo-beta-lactamase-
M AN U
producing carbapenem-resistant Escherichia coli associated with exposure to duodenoscopes. Jama 2014;312:1447-55. 4.
Wendorf KA, Kay M, Baliga C, et al. Endoscopic Retrograde Cholangiopancreatography-Associated AmpC Escherichia coli Outbreak. Infect Control Hosp Epidemiol 2015;Epub ahead of print (3/30/2015):1-9. Alrabaa S. Early identification and control of carbapenemase-producing
TE D
5.
Klebsiella pneumoniae, originating from contaminated endoscopic equipment.
6.
EP
Am J Infect Control 2013;41:850.
Verfaillie CJ, Bruno MJ, A FVItH, et al. Withdrawal of a novel-design
AC C
duodenoscope ends outbreak of a VIM-2-producing Pseudomonas aeruginosa. Endoscopy 2015;Epub ahead of print (3/31/2015).
7.
U.S. Food and Drug Administration. Design of Endoscopic Retrograde Cholangiopancreatography (ERCP) Duodenoscopes May Impede Effective Cleaning: FDA Safety Communication, 2015.
8.
Petersen BT, Koch J, Ginsberg GG. Infection Using ERCP Endoscopes. Gastroenterology 2016;151:46-50. 23
ACCEPTED MANUSCRIPT
9.
Smith ZL, Oh YS, Saeian K, et al. Transmission of carbapenem-resistant Enterobacteriaceae during ERCP: time to revisit the current reprocessing guidelines. Gastrointest Endosc 2015;81:1041-5. Rutala WA, Weber DJ. Outbreaks of carbapenem-resistant Enterobacteriaceae
RI PT
10.
infections associated with duodenoscopes: What can we do to prevent infections? Am J Infect Control 2016;44:e47-51.
Centers for Disease Control and Prevention. Interim Duodenoscope Culture
SC
11.
Method, 2015.
Chen A LB, Wright SB, Yassa DS, Snyder GM, Hirsch EB. Delayed processing of
M AN U
12.
simulated duodenoscope and patient cultures yields viable bacteria. Society for Healthcare Epidemiology of America Spring 2016 Conference. Atlanta, Georgia, U.S.A., 2016.
Centers for Disease Control and Prevention. Interim Duodenoscope Surveillance
TE D
13.
Protocol, 2015 Available at: http://www.cdc.gov/hai/organisms/cre/creduodenoscope-surveillance-protocol.html. Accessed 1/5/2017 . Ross AS, Baliga C, Verma P, et al. A quarantine process for the resolution of
EP
14.
duodenoscope-associated transmission of multidrug-resistant Escherichia coli.
15.
AC C
Gastrointest Endosc 2015;82:477-83.
Paula H, Presterl E, Tribl B, et al. Microbiologic surveillance of duodenoscope reprocessing at the Vienna University Hospital from November 2004 through March 2015. Infect Control Hosp Epidemiol 2015;36:1233-5.
24
ACCEPTED MANUSCRIPT
16.
Brandabur JJ, Leggett JE, Wang L, et al. Surveillance of guideline practices for duodenoscope and linear echoendoscope reprocessing in a large healthcare system. Gastrointest Endosc 2016;84:392-399 e3. Naryzhny I, Silas D, Chi K. Impact of ethylene oxide gas sterilization of
RI PT
17.
duodenoscopes after a carbapenem-resistant Enterobacteriaceae outbreak. Gastrointest Endosc 2016;84:259-62.
Almario CV, May FP, Shaheen NJ, et al. Cost Utility of Competing Strategies to
SC
18.
Prevent Endoscopic Transmission of Carbapenem-Resistant
19.
M AN U
Enterobacteriaceae. Am J Gastroenterol 2015;110:1666-74.
Chen A, Lane BV, Wright SB, Yassa DS, Snyder GM, Hirsch EB. Delayed processing of simulated duodenoscope and patient cultures yields viable bacteria. Poster presentation, Society for Healthcare Epidemiology of America
20.
TE D
Spring 2016 Conference, Atlanta, GA May 18-21, 2016).
https://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/M
EP
edicalDevices/MedicalDevicesAdvisoryCommittee/GastroenterologyUrologyDevicesPanel/ucm445590.htm. accessed 5/27/17. Rutala WA, Weber DJ. Outbreaks of carbapenem-resistant Enterobacteriaceae
AC C
21.
infections associated with duodenoscopes: What can we do to prevent infections? Am J Infect Control 2016; 44(5 Suppl): e47-51
22.
Heroux R, Sheppard M, Wright SB, Sawhney M, Hirsch EB, Kalaidjian R, Snyder GM. Duodenoscope hang time does not correlate with risk of bacterial contamination. Am J Infect Control. 2016 Dec 26. PMID: 28034537
25
ACCEPTED MANUSCRIPT
23.
Vickery K, Ngo QD, Zou J, Cossart YE. The effect of multiple cycles of contamination, detergent washing, and disinfection on the development of biofilm in endoscope tubing. Am J Infect Control. 2009;37(6):470-5. Pajkos A, Vickery
RI PT
K, Cossart Y. Is biofilm accumulation on endoscope tubing a contributor to the failure of cleaning and decontamination? J Hosp Infect. 2004;58(3):224-9 24.
Pajkos A, Vickery K, Cossart Y. Is biofilm accumulation on endoscope tubing a
SC
contributor to the failure of cleaning and decontamination? J Hosp Infect. 2004;58(3):224-9.
Olafsdottir LB, Wright SB, Smithey A, Herous R, Hirsch EB, Chen A, Lane B,
M AN U
25.
Sawhney MS, Snyder GM. Adenosine Triphosphate Quantification Correlates Poorly with Microbial Contamination of Duodenoscopes. Infect Control Hosp Epidemiol, 2017;17:1-7
Sethi S, Huang RJ, Barakat MT, Banaei N, Friedland S, Banerjee S. Adenosine
TE D
26.
triphosphate bioluminescence for bacteriologic surveillance and reprocessing strategies for minimizing risk of infection transmission by duodenoscopes.
27.
EP
Gastrointest Endosc. 2017 Jun;85(6):1180-1187. https://www.fda.gov/MedicalDevices/Safety/AlertsandNotices/ucm454766.htm.
28.
AC C
Accessed 6/22/17.
Rutala WA, Weber DJ. ERCP scopes: what can we do to prevent infections? Infect Control Hosp Epidemiol. 2015 Jun;36(6):643-8.
26
ACCEPTED MANUSCRIPT
FIGURE 1: Frequency of no growth (0 CFU), low quantity of growth (1-9 CFU) and significant quantity of growth (≥10 CFU) in each of the trial arms Note: CFU, colony-forming units; ETO, ethylene oxide gas sterilization; HLD, high-level
RI PT
disinfection
FIGURE 2: Culture results from individual study duodenoscopes for each day
SC
sampled during the study period.
Note: CFU, colony-forming units; dHLD, double high-level disinfection; HLD/ETO, high-
M AN U
level disinfection/ethylene oxide gas sterilization; sHLD, standard high-level disinfection. The x-axis represents all dates during the study period for which one or more duodenoscopes were cultured; the y-axis represents the 18 (de-identified) duodenoscopes. Empty cells indicate no culture was taken for that duodenoscope on
TE D
that day. The number in each cell represents the highest CFU of aerobic bacterial growth from either the elevator mechanism or working channel, including in cases
AC C
EP
where a duodenoscope was cultured more than once on a given day.
27
ACCEPTED MANUSCRIPT
TABLE 1: Frequency of the primary outcome (≥1 multidrug-resistant organism), or secondary outcomes of any growth > 0 CFU and growth of ≥10 CFU on any duodenoscope culture
RI PT
Growth, elevator mechanism or working channel (%) (N)
≥1 MDRO
>0 CFU*
≥10 CFU†
sHLD
174
0
28 (16.1)
4 (2.3)
dHLD
169
0
27 (16.0)
7 (4.1)
HLD/ETO 173
0
39 (22.5)
0
94 (18.3)
516
9 (4.2)
20 (3.9)
M AN U
Total
SC
Trial Arm
Note: CFU, colony-forming units; dHLD, double high-level disinfection; HLD/ETO, standard high-level disinfection followed by ethylene oxide gas sterilization; MDRO, multidrug-resistant organism; sHLD, standard high-level disinfection.
TE D
See Methods and the Supplementary Material for definition of MDRO.
AC C
EP
* p = 0.21, † p = 0.36 by Fisher’s exact test
ACCEPTED MANUSCRIPT
TABLE 2: Association between trial arm and duodenoscope contamination in generalized estimating equations regression models Duodenoscope
contamination with >0 CFU
contamination with ≥10 CFU
Variable
RR (95%CI)
sHLD
ref
p-value
Unadjusted
RR (95%CI)
p-value
ref
SC
—
model
RI PT
Duodenoscope
—
1.02 (0.66-1.58)
0.92
2.05 (1.06-3.94)
0.03
HLD/ETO
1.43 (0.89-2.29)
0.14
2.40 (1.13-5.09)
0.02
Adjusted sHLD
ref
model*
M AN U
dHLD
—
—
—
1.02 (0.66-1.58)
0.93
— —
—
HLD/ETO
1.41 (0.86-2.32)
0.17
— —
—
TE D
dHLD
Note: 95%CI, 95% confidence interval; CFU, colony-forming units; sHLD, standard highlevel disinfection; dHLD, double high-level disinfection; HLD/ETO, high-level disinfection
EP
followed by ethylene oxide gas sterilization; RR, relative risk. The adjusted model included duration of preceding procedure (in hours), use of sphincterotomy and stent
AC C
removal during the preceding procedure, and time elapsed between preceding procedure and repeat culture (in days), all non-significant (full regression models are presented in the Supplementary Material).
ACCEPTED MANUSCRIPT
TABLE 3: Characteristics of patients enrolled in study. Patients undergoing procedures with a duodenoscope assigned to a study arm
N, (%)
All patients
RI PT
Characteristic
HLD/ETO (N=158)
dHLD
sHLD
(N=166)
(N=171)
90 (54.2)
81 (47.4)
66.5 (16.3)
63.0 (17.5)
62.8 (15.9)
27.7 (6.9)
27.9 (6.5)
27.3 (5.5)
19 (12.0)
26 (15.7)
14 (8.2)
(N=495)*
SC
Patient characteristics 236 (47.7)
Age, mean, years (SD)
64.0 (16.6)
Body mass index, mean, kg/m2 (SD)
27.6 (6.3)
Hepatic or pancreaticobiliary mass
59 (11.9)
Biliary abnormality†
181 (36.6)
60 (38.0)
66 (39.8)
55 (32.2)
Cholangitis
46 (9.3)
13 (8.2)
9 (5.4)
24 (14.0)
Solid organ transplant‡
16 (3.2)
6 (3.8)
6 (3.6)
4 (2.3)
Hematologic malignancy
6 (1.2)
0
1 (0.6)
5 (2.9)
124 (25.1)
41 (26.0)
39 (23.5)
44 (25.7)
5 (1.0)
2 (1.3)
1 (0.6)
2 (1.2)
Chronic hemodialysis Procedure characteristics
TE D
EP
AC C
Diabetes mellitus, type 1 or 2
65 (41.1)
M AN U
Female gender
ACCEPTED MANUSCRIPT
Procedure duration, mean, minutes
32.9 (18.9)
33.7 (20.8)
Sphincterotomy
208 (42.0)
67 (42.4)
Stent placement
187 (37.8)
61 (38.6)
Stent removal
132 (26.7)
41 (25.6)
Balloon dilation or stone extraction
301 (60.8)
99 (62.7)
Brushing for cytology
71 (14.3)
32.4 (18.4)
32.7 (17.6)
M AN U
SC
RI PT
(SD)
18 (11.4)
Full demographic and procedure data unavailable for 3 of 498 patients
†
Biliary duct obstruction, stenosis, or leak
‡
Liver (14), kidney (3); one patient had both a liver and kidney transplant
TE D
*
65 (39.2)
76 (44.4)
60 (36.1)
66 (38.6)
46 (27.7)
45 (26.3)
104 (62.7)
98 (57.3)
24 (14.5)
29 (17.0)
Note: Patients undergoing multiple procedures are included for each procedure that was performed during the study
AC C
EP
period. See text for details regarding frequency of repeat procedures.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT