Keywords Quality management - Hygiene - Quality and logistical aspects
Introduction
Endoscope disinfection is a necessary component of safe endoscopy practice. However,
endoscope reprocessing is a complex, tedious, multistep process that includes pre-cleaning,
manual cleaning, high-level disinfection (HLD), rinsing and drying, and storage. Lapses
in reprocessing protocols are identified in the vast majority of endoscope-related
infection outbreaks [1 ].
Methods to ensure adequate reprocessing are currently lacking. Surveillance bacterial
cultures have been proposed for this purpose but are limited by inherently delayed
results that are not immediately actionable for endoscopes inadequately reprocessed
[2 ]. Point-of care testing for adenosine triphosphate (ATP), a marker of bioburden,
may be a reasonable surrogate to assess the effectiveness of manual cleaning, which
is critical for subsequent HLD but also the most prone to error. ATP testing has recently
been shown to correlate with quality of endoscope cleaning [3 ], and ATP values <200 relative light units (RLUs) after manual cleaning of endoscopes
and have been associated with subsequent adequate disinfection [4 ]
[5 ].
The World Gastroenterology Organization (WGO) sponsors training centers that train
digestive health professionals internationally. Many of these training centers are
in limited resource regions, potentially impacting endoscope reprocessing capabilities.
We sought to evaluate endoscope reprocessing across multiple WGO training centers
using ATP testing, before and after optimization of reprocessing methods.
Methods
Setting
This was a multicenter study involving five WGO training centers that volunteered
for study participation and met enrollment criteria: Bangkok, Thailand; Bogota, Colombia;
San Jose, Costa Rica; Suva, Fiji; Nairobi, Kenya. A two-person study team consisting
of one nurse (of 2 nurses total) with experience in performing and supervising endoscope
reprocessing, and one physician (of 3 physicians total) with knowledge about reprocessing
guidelines visited each site for 1 week. The study was conducted between May 2018
and February 2020. Two additional training centers planned to participate in the study,
but COVID-19 pandemic-induced changes in endoscopy center practices and case volumes,
as well as limitations on international travel, prevented these sites from participating.
Prior to and upon arrival at the study site, the study protocol ([Fig. 1 ]) was reviewed with local training center team members including physicians, nurses,
and reprocessing personnel. Consecutive endoscopes used in patients who gave consent
for endoscopic procedures were included for investigation. Baseline reprocessing practices
were observed to understand workflow and study integration (0.5 days). Subsequently,
baseline ATP testing was performed during the first half of the site visit (2 days).
Following this period, a meeting was held among study staff and local team members
to review baseline data and potential interventions to optimize reprocessing practices
(0.5 days). Following the implementation of immediately feasible interventions, post-intervention
ATP testing was repeated during the latter half of the visit (2 days). At one center,
examination of endoscope suction/biopsy channels was performed with a borescope (Clarus
Medical LLC, Minneapolis, Minnesota, United States).
Fig. 1 Study flow.
Endoscopic procedures including upper endoscopy, flexible sigmoidoscopy, colonoscopy,
endoscopic retrograde cholangiopancreatography, and upper endoscopic ultrasound were
performed by each training center’s endoscopy faculty and trainees using existing
endoscopes. Endoscope data (manufacturer and model number, identification [ID] number,
number of procedure/reprocessing cycles, and repair history), when available, were
collected from each site. For each reprocessing cycle investigated, the endoscope
ID number, the type of procedure (upper/lower), whether an endoscopic intervention
(e.g., forceps biopsy) was performed, whether pre-cleaning was performed, reprocessing
technician ID, as well as ATP results were recorded by study staff. No clinical information
was collected.
Endoscope reprocessing was expected to adhere to existing manufacturer and WGO/World
Endoscopy Organization (WEO) global guidelines [6 ]. This includes immediate bedside pre-cleaning with wiping of the exterior surface
to remove visible debris and flushing of the suction/biopsy channel with 250 mL enzymatic
cleaning solution suctioned through the endoscope. This is followed by endoscope transportation
to a dedicated reprocessing room for manual cleaning including leak testing, brushing
of internal channels and components, and irrigation of detergent and water through
the channels. HLD is then performed, either manually or by automated endoscope reprocessors.
The endoscopes are then dried and if not immediately reused, hung in a storage cabinet.
ATP testing
ATP testing was performed on rinsates from the endoscope suction/biopsy channel at
four junctures of reprocessing: 1) before manual cleaning; 2) after manual cleaning;
3) after HLD; and 4) after overnight storage (for endoscopes immediately reused after
HLD, after overnight storage testing was not possible for that specific reprocessing
cycle). ATP levels were assessed using Clean-Trace Water ATP tests and a compatible
luminometer (3M, St. Paul, Minnesota, United States) quantifying ATP bioluminescence
in relative light units ([Fig. 2 ]). Prior validation studies for this specific ATP assessment kit support the post-manual
cleaning benchmark of <200 RLUs used in this study [4 ]
[5 ]. ATP testing was performed by first purging the channel with an air-filled syringe
to clear any gross residual channel contents. Forty milliliters of local tap water
was then flushed through the channel followed by air, and the effluent was collected
in a disposable cup. Background ATP measurements of the cup containing 40 mL of locally
available tap water (prior to any contact with an endoscope) were performed to assess
for significant confounding, and showed negligible values of 3 to 11 RLUs. The training
center team, including reprocessing personnel, was blinded to baseline ATP testing
results in an effort avoid impacting baseline reprocessing practices.
Fig. 2 ATP sampling protocol. The suction/biopsy channel is purged with air and then flushed
with 40 mL of water. The channel is then purged again and the effluent is collected
in a disposable cup. The distal portion of the ATP stick is submerged in the sample,
and then returned to its holder and depressed to activate the bioluminescence reaction.
This reaction is then quantified by the luminometer in the form of relative light
units (RLUs).
Intervention
Following baseline ATP testing, a meeting was held among the study staff and local
training center stakeholders including physicians, nurses, and reprocessing personnel.
Baseline ATP testing data were reviewed in the context of existing reprocessing practices
and reprocessing guidelines. Opportunities for improvement in each phase of the reprocessing
protocol (pre-cleaning, cleaning, HLD, rinsing and drying, and storage) were identified
collaboratively by training center members and study staff. Modifications of local
reprocessing practices were determined according to WGO/WEO global guidelines for
endoscope disinfection [6 ]. Specifically, these were based on Cascade options, which outlines a hierarchy of
standard procedures that allow for alternatives in resource-sensitive steps in endoscope
reprocessing, particularly in areas of the world in which external factors limit available
options. For example, in a training center in which renewing the cleaning detergent
solution for each new procedure (“medium-extensive” resources level) was not possible,
renewal at an interval that the center’s resources allowed was permitted (“limited”
resources level). After implementing the maximum feasible interventions in consensus
with local stakeholders, follow-up ATP testing was performed in similar fashion to
baseline ATP testing except training center staff (including reprocessing personnel)
were no longer blinded to ATP results. Of note, endoscopes with a post-manual cleaning
ATP ≥200 RLUs (pre- or post-intervention) were not triaged individually to any specific
intervention (e.g., repeat manual cleaning or cycle of reprocessing), and were permitted
to proceed with the existing reprocessing protocol.
Interventions in reprocessing practices that were implemented at each study site are
detailed in [Table 1 ]. These most commonly included the performance or optimization of bedside pre-cleaning,
more frequent exchange or increased volumes of manual cleaning solution, longer rinsing
times, and longer drying times with vertical storage after HLD in accordance with
manufacturer recommendations and/or WGO/WEO global guidelines. At Site D, staff had
recently been trained in endoscope reprocessing, and no suggestions for improvement
were identified.
Table 1 Reprocessing interventions performed at each site.
Study site
Potential for optimization
Intervention
A
Pre-cleaning:
Suctioned fluid by time, not volume of enzymatic solution
Duodenoscope/echoendoscope elevator was stationary while suctioning water (open or
closed)
The air/water cleaning button was not used during pre-cleaning
Delays after pre-cleaning and before leak testing and manual cleaning
Final Rinsing:
Pre-cleaning:
Standardized suctioning of 250 mL enzymatic solution
Toggle duodenoscope/echoendoscope elevator up and down while suctioning water
Use the air/water cleaning button during pre-cleaning
Minimize delays after pre-cleaning and before leak testing and manual cleaning
Final Rinsing:
B
Pre-cleaning:
Frequent delays in pre-cleaning
Inadequate volume of enzymatic solution (approximately 50 ml)
The air/water cleaning button was not used during pre-cleaning
Water jet channels not pre-cleaned
Dirty scopes hung temporarily on the endoscope tower alongside clean scopes
Manual cleaning:
Leak testing not performed prior to scope immersion in liquid
Gloves used to plug the sink left in place overnight
Water in the cleaning sink is cold, preventing activation of enzymes in the enzymatic
detergent solution
All endoscope channels not brushed during cleaning
Biopsy ports not adequately brushed
HLD
Final rinsing
Drying and Storage
Air is applied briefly, and borescope exam shows retained water in the suction/biopsy
channel
Scopes sometimes stored with valves and caps on
Scope umbilicus stored in a “U” configuration, facing up
Accessories
Pre-cleaning:
Pre-cleaning performed immediately after endoscope withdrawal from the patient, by
the endoscopist if other staff are busy
Increase bedside pre-cleaning volume to 250 mL enzymatic solution
Use the air/water cleaning button during pre-cleaning
For scopes with water jet channels, flush by activating the foot pedal or using a
syringe
Do not bring clean scopes into the endoscopy room until dirty scopes have been taken
to the cleaning room
Manual cleaning:
Perform leak testing before immersing scope in liquid
All devices in the cleaning sink discarded, changed or disinfected at the end of each
day
Use water >20°C in the cleaning sink
Brush the entire suction/biopsy channel via the suction valve port (2 directions)
and also through the biopsy port
A short, large caliber brush (“stubby brush”) should be used
HLD
Final Rinsing
Drying and Storage
Longer duration drying (10 minutes) using compressed air, per manufacturer’s recommendations
Do not leave valves and caps on scopes while in storage
Both the insertion shaft and the umbilicus should be stored hanging downward
Accessories
C
Pre-cleaning:
Manual cleaning:
Occasionally delays to initiation of manual cleaning
Insufficient volume of enzymatic in sink to allow complete endoscope immersion
Enzymatic recycled and replaced daily
HLD:
Drying and Storage:
Pre-cleaning:
Manual cleaning:
Minimize delays to initiation of manual cleaning
Increase volume of enzymatic in sink to allow complete endoscope immersion
Replace enzymatic solution every 3 endoscopes (ideally replace with each endoscope)
HLD:
Drying and Storage:
D
No suggestions for improvement
No suggestions for improvement
E
Pre-cleaning:
Manual cleaning:
Delays between pre-cleaning and manual cleaning
Rinse with recycled water
Residue on scope exterior prior to HLD
Gloves not being changed when moving from dirty to clean areas
HLD:
Pockets of air when flushing through channels
Imprecise HLD dwell times
Rinse with recycled water
Drying and Storage:
Pre-cleaning:
Manual cleaning:
Minimize/eliminate delays between pre-cleaning and manual cleaning
Rinse with fresh water
Remove detergent residue from scope exterior
Change gloves every time moving from dirty to clean areas
HLD:
Avoid pockets of air when flushing solution through channels
Use of a timer to ensure appropriate dwell times
Continuous supply of fresh water for rinsing
Drying and Storage:
Endpoints and sample size
The primary endpoint of this study was comparison of mean ATP levels after manual
cleaning between baseline and post-intervention results among all participating sites.
To detect a difference of ≥30% in mean ATP levels after intervention with 90% power,
a total of 200 pre- and 200 post-intervention endoscopes were needed across all centers.
To detect a difference of ≥30% with 70% power at each individual site, 26 pre- and
26 post-intervention endoscopes were needed per site; this number of reprocessing
cycles was not achieved pre-intervention at one of the five study sites. Secondary
endpoints included changes in mean ATP measurements at all other time points after
intervention, and analysis of the association between the study variables and ATP
measurements.
Statistical analysis
ATP levels were compared pre- and post-intervention using the Pearson chi-squared
and Kruskal-Wallis tests. Differences in RLU measurements before and after the intervention
were formally analyzed using a hierarchical model. ATP levels were log-transformed
to mitigate the effect of right-skewed data on the model estimates. We expected substantial
variability in the pre-intervention results and intervention effect from site to site,
and thus modeled separate site-effects (for intercept and treatment) as random effects
in the hierarchical model. The estimate of the overall fixed treatment effect (with
95% confidence interval [CI] and P value) served as the primary estimate for each model. The treatment effect for post-manual
cleaning time point served as the primary endpoint. Other time points were analyzed
as secondary endpoints. Additionally, we calculated the 95% confidence interval for
the proportion of post-intervention, post-manual cleaning ATPs which are below the
recommended benchmark of 200 RLUs.
Results
Among the five study sites, a total of 343 endoscope reprocessing cycles were studied
(160 pre-intervention and 183 post-intervention) and a total of 1182 ATP tests were
performed ([Table 2 ]). Sixty-five endoscopes were studied. Many had been acquired used, and data on the
date of manufacture, number of uses, and date of most recent servicing by the manufacturer
were not available for a substantial number of endoscopes. For endoscopes with available
data, mean scope age was 5.3 years (range 1–13, N=38), and mean time since last servicing
was 2.0 years (range 0.1–8, N=46).
Table 2 Total, pre-intervention and post-intervention ATP measurements.
Total
Pre-intervention
Post-intervention
P value
ATP, adenosine triphosphate; Q, quarter; RLU, relative light unit; HLD, high-level
disinfction
No. reprocessing cycles studied per site N=343
0.84
A
102 (29.7%)
52 (32.5%)
50 (27.3%)
B
75 (21.9%)
34 (21.2%)
41 (22.4%)
C
58 (16.9%)
27 (16.9%)
31 (16.9%)
D
46 (13.4%)
19 (11.9%)
27 (14.8%)
E
62 (18.1%)
28 (17.5%)
34 (18.6%)
Pre-clean performed N=342
<0.001
No
28 (8.2%)
28 (17.6%)
0 (0.0%)
Yes
314 (91.8%)
131 (82.4%)
183 (100.0%)
Type of procedure N=333
0.22
130 (39.0%)
55 (35.5%)
75 (42.1%)
203 (61.0%)
100 (64.5%)
103 (57.9%)
Type of scope N=343
0.68
124 (36.2%)
56 (35.0%)
68 (37.2%)
219 (63.8%)
104 (65.0%)
115 (62.8%)
Before manual cleaning ATP (RLUs) N=337
<0.001
Median (Q1, Q3)
1343 (352, 5403)
2709 (540, 12491)
760 (257, 2346)
Range
15–739650
51–739650
15–44145
After manual cleaning ATP (RLUs) N=342
0.23
Median (Q1, Q3)
52 (20, 134)
56 (24, 158)
51 (18, 122)
Range
2–6760
2–2840
3–6760
After manual cleaning ATP (RLUs) N=342
0.11
≥200
61 (17.8%)
34 (21.4%)
27 (14.8%)
<200
281 (82.2%)
125 (78.6%)
156 (85.2%)
After HLD ATP (RLUs) N=333
0.54
Median (Q1, Q3)
26 (10, 69)
26 (11, 69)
27 (9, 70)
Range
2–1473
2–513
2–1473
Before clinical use ATP (RLUs) N=171
0.72
Median (Q1, Q3)
44 (12, 162)
41 (8, 165)
44 (16, 161)
Range
1–7389
1–7389
1–4313
The majority of procedures with gastroscopes were for upper endoscopy (63.8%) and
involved an endoscopic procedural intervention requiring passage of a device via the
suction port (61.0%), most commonly biopsies. Only 4.0% of procedures were endoscopic
retrograde cholangiopancreatographies or endoscopic ultrasound, requiring the use
of an endoscope with an elevator. Total ATP values and values for individual sites
are shown in [Table 2 ]. Pre-intervention, 34 (21.4%) reprocessing cycles had an ATP measurement greater
than the threshold of 200 RLUs following manual cleaning.
Interventions
Pre- and post-intervention procedures were similar in the proportions of diagnostic-only
procedures as well as the type of scope used ([Table 2 ]). Each site performed a similar number of procedures before and after the intervention.
Pre- and post-intervention ATP data from individual sites are shown in Supplemental Fig. 1 . Intervention resulted in a dramatic decrease in pre-manual cleaning ATP values at
one site that had not been consistently performing bedside pre-cleaning prior to the
study intervention (Site C, median ATP 158,807 RLUs pre-intervention, and 1,730 RLUs
post-intervention, P <0.001). However, there were no meaningful changes in ATP values by site after manual
cleaning, or at any later stage in the reprocessing cycle.
Considering all sites together, there were significant improvements in the performance
of bedside pre-cleaning from 82.4% to 100% (P <0.001) as well as in median ATP values before manual cleaning. This difference in
pre-manual cleaning ATP values remained significant after removing Site C (median
ATP 1951 RLUs vs. 603 RLUs, P <0.001). However, there was no significant reduction in ATP values after manual cleaning
or at later stages of the reprocessing cycle or after overnight storage ([Table 1 ]). The relative post-intervention reduction in mean ATP values after manual cleaning
was 19% (P = 0.17; 95% CI: 40% reduction, 10% increase). There was a post-intervention downward
trend in the proportion of endoscopes with ATP ≥200 RLUs after manual cleaning, which
was not statistically significant (21.4% vs. 14.8%, P =0.11, odds ratio [OR] 0.64; 95% CI 0.36–1.11). Mixed-effect modeling, recognizing
that measures within each site may be correlated, yielded similar results (analysis
not shown).
Associations with post-manual cleaning ATP levels below 200 were inspected separately
before and after the educational intervention with study site, type of endoscope (upper
vs. lower), and endoscopic procedure intervention ([Table 3 ]). Colonoscopes consistently demonstrated lower ATP values than gastroscopes. Across
pre- and post-intervention observations, post-manual cleaning ATP values were ≥200
RLUs in 1.6% of colonoscope reprocessing cycles vs. 27.1% of gastroscope reprocessing
cycles (P <0.001). Post-manual cleaning ATP levels were <200 RLU in 80.8% endoscopes undergoing
manual HLD versus 84.1% endoscopes undergoing automated HLD (P =0.47).
Table 3 Unadjusted comparison of associations with post-manual cleaning ATP < 200
RLUs.
Pre-intervention
Post-intervention
% ATP <200 RLUs
P value
% ATP <200 RLUs
P value
*Missing data: There was 1 pre-intervention observation with pre-cleaned
performed unknown (with post manual cleaning [PMC] ATP < 200 RLU). There were 5
pre-intervention observations and 5 post-intervention observations with endoscopic
intervention unknown (all with PMC-ATP < 200 RLUs). There were 6
post-intervention observations with pre-manual cleaning ATP unknown (5 with PMC-ATP
< 200 RLUs). There were 8 post-intervention observations with after HLD ATP
unknown (6 with PMC-ATP < 200 RLUs). There were 74 pre-intervention observations
(58 with PMC-ATP < 200 RLUs) and 98 post-intervention observations (83 with
PMC-ATP < 200 RLUs) with before clinical use ATP unknown.
ATP, adenosine triphosphate; RLU, relative light unit; PMC, ; Q, quarter; HLD,
high-level disinfection.
Site
A
82% (42/51)
0.62
90% (45/50)
0.25
B
71% (24/34)
90% (37/41)
C
74% (20/27)
87% (27/31)
D
84% (16/19)
74% (20/27)
E
82% (23/28)
79% (27/34)
Pre-clean performed*
No
75% (21/28)
0.62
NA (0/0)
NA
Yes
79% (103/130)
85% (156/183)
Any endoscopic intervention*
Diagnostic only
87% (48/55)
0.037
88% (66/75)
0.32
Any intervention done
73% (72/99)
83% (85/103)
Type of scope
Lower
98% (55/56)
<0.001
99% (67/68)
<0.001
Upper
68% (70/103)
77% (89/115)
Median (Q1, Q3)
P value
Median (Q1, Q3)
P value
Before manual cleaning ATP (RLUs)*
Post-manual ATP <200 RLUs
1866 (437, 8773)
<0.001
518 (206, 1584)
<0.001
Post-manual ATP ≥200 RLUs
11916 (3620, 27964)
5433 (1862, 8422)
After HLD ATP (RLUs)*
Post-manual ATP <200 RLUs
20 (8, 45)
<0.001
20 (8, 48)
<0.001
Post-manual ATP ≥ 200 RLUs
78 (42, 183)
91 (54, 182)
Before clinical use ATP (RLUs)*
Post-manual ATP <200 RLUs
22 (6, 114)
<0.001
42 (12, 129)
0.011
Post-manual ATP ≥200 RLUs
170 (72, 867)
148 (38, 408)
In multivariable logistic modeling, gastroscopes were significantly less likely (OR
0.04, 95% CI 0.01, 0.19; P <0.001) than colonoscopes to achieve post-manual cleaning ATP <200 RLU. No other
factor (educational intervention, study site, endoscope age) was significantly associated
with improved cleaning outcomes. Sites which performed manual versus automated HLD
did not have significantly different likelihood of post-manual cleaning ATP <200 RLU
(OR 1.18, 95% CI 0.56–2.50; P =0.67).
Suction/biopsy channels of 12 endoscopes were examined with a borescope at Site B,
and findings were rated using a previously described scale [7 ]. Total scores ranged from 5 to 16, with channel scratches observed in all endoscopes,
adherent peel in four, buckling in one, and channel perforation in one. Scores did
not correlate with mean post-manual cleaning ATP levels (correlation coefficient 0.32).
Discussion
Endoscope reprocessing is a critical component in the safe performance of endoscopic
procedures, but is inherently challenging and prone to error. Recent endoscope-related
infection outbreaks have obligated endoscopy units to reevaluate their overall reprocessing
practices. We observed baseline reprocessing practices at five international WGO training
centers in resource-limited settings and assessed the impact of interventions on reprocessing
quality, as measured by ATP levels. Several interventions were made, aligning reprocessing
procedures with best practices and resulting in a downward trend in the proportion
of endoscopes with post-manual cleaning ATP above the benchmark of 200 RLUs from 21%
to 14%, a difference that was not statistically significant. Our findings suggest
that a significant proportion of endoscope reprocessing cycles may not result in adequate
endoscope disinfection, even after expert review and optimization of local practices.
In an effort to reduce the risk of endoscope-related infection transmission, several
modalities for assessing reprocessing quality assurance have either been recommended
or proposed. However, most are associated with significant barriers, especially in
limited resource settings. For example, some guidelines recommend routine microbiological
surveillance of endoscopes and reprocessing equipment with bacterial cultures which
carry high sensitivity for microbial contamination but are inherently limited by results
that require over 1 to 2 days to become available and are not immediately actionable
for endoscopes that may be inadequately reprocessed [2 ]. For duodenoscopes specifically, recent recommendations to acquire partially or
fully disposable endoscopes is often not economically feasible in resource-constrained
contexts. This underscores the importance of optimizing existing reprocessing practices.
Bioluminescent testing for ATP in biologic residue has been used for quality reassurance
in the food service industry and recently applied to the healthcare setting. ATP testing
is relatively easy to perform and provides real-time results. Alfa and colleagues
established a post-manual cleaning benchmark of <200 RLUs of the suction/biopsy and
air/water channels (specific to the commercially available kit used in this study,
but using sterile water) in simulated use settings [4 ]
[5 ]. This benchmark has been proposed as an indicator of adequate cleaning prior to
submitting the endoscope for HLD. Manual cleaning remains perhaps the most critical
step in the reprocessing cycle, because biofilm can form when contaminated endoscopes
undergo repeated cycles of reprocessing, creating a protective matrix allowing viable
organisms to survive HLD. Subsequent disinfection or sterilizing processes can fail
if the instrument has not been sufficiently cleaned.
Despite identifying and addressing gaps throughout the reprocessing cycles, we did
not find a satisfactory reduction in the post-manual cleaning ATP level. We suspect
this is multifactorial in etiology. For example, upper but not lower endoscopes were
associated with greater post-manual cleaning ATP levels, an observation made previously
and possibly related to a narrower or more damaged suction/biopsy channel that renders
manual cleaning more challenging [8 ]. Although reprocessing practices were optimized, this was still limited by training
center-specific availability of resources, so manufacturer reprocessing recommendations
may not have been fully met. For example, in an effort to conserve pre-cleaning enzymatic
solution, one training center was exchanging solution daily as opposed to with every
scope. In discussion with training center stakeholders it was decided to exchange
solution every third endoscope. It also is unknown if the persistent proportion of
endoscopes with elevated ATP level can be further reduced by a second consecutive
round of manual cleaning (or HLD), which has been variably successful in other studies
[3 ]
[9 ].
We attempted to determine the number of previous procedures as well as repair and
servicing history for each endoscope but this information was not available for many
endoscopes. Defects in an endoscope combined with inadequate reprocessing (including
drying and storage) increase the potential for biofilm formation. Although not part
of the study design a priori, we explored the use of a thin fiberoptic borescope at
one study site to pass through and inspect the suction/biopsy channel for damage that
may predispose to biofilm formation and/or interfere with reprocessing [7 ]. All examinations were performed by a single study team member and assessed using
a previously described scoring system. A median damage score 6.5 (range 2–14) and
median total score of 9.5 (range 5–16) were calculated with notable findings including
debrided area of channel and one potential channel perforation, although the significance
of these findings and their correlation with ATP levels is unclear at this time and
remain an area of future investigation.
The complicated, multistep, tedious nature of endoscope reprocessing lends itself
to human error and so-called human factors and observations during our study are supportive
of this. In a recent large survey of US and international health care workers on endoscope
reprocessing, 70% reported feeling pressured to work more quickly and 17% admitted
to skipping steps or performing them more quickly [10 ]. In our study we identified multiple sites that deferred bedside pre-cleaning or
performed it with suboptimal volume. While this and other omissions may be related
to pressures of endoscope/procedure turnaround time, this is likely to be compounded
in limited resource settings where the limited number of available endoscopes, shortages
and cost of reprocessing supplies, and training and competency testing may be less
favorable. With education of training center medical team and reprocessing staff,
pre-cleaning in this study improved to 100%, and study staff observed adherence to
study team recommendations post-intervention, but results still fell short of expectations.
This suggests that the site-specific optimized process of endoscope reprocessing is
not adequate to achieve desired levels of disinfection in some cases.
We found ATP testing to be feasible, informative for guiding reprocessing interventions,
and well received by reprocessing personnel for quality assurance. Enrolling consecutively
used endoscopes and testing them after each phase of reprocessing did not appear to
significantly impede procedure workflow. Because collection and testing of samples
for ATP is rapid (1–2 minutes), is most valuable after the manual cleaning phase,
and may only be necessary for periodic surveillance, a surveillance program is unlikely
to interfere with typical workflow. Reprocessing staff generally were engaged throughout
the study and at the time of intervention were collaborative in identifying potential
sources of residual bioburden and solutions to address them. The study visit and intervention
also were used as an opportunity for reprocessing staff to inquire about preexisting
reprocessing questions, because many reprocessing staff had not received formal training
or were uncomfortable interpreting manufacturer reprocessing instructions. In the
post-intervention period, reprocessing staff were eager to understand the impact of
the various interventions on reprocessing quality as assessed by ATP. Quantitative
ATP RLUs provided instant feedback, and values meeting the benchmark resulted in positive
reinforcement. Moreover, reprocessing staff were taught to perform ATP testing in
the post-intervention period and demonstrated competency in doing so, further supporting
its ease of adoption.
Our study has some notable limitations. Although this was a multicenter, prospective
study, baseline reprocessing techniques were not standardized across all institutions.
For example, there were variations in reprocessing techniques, supplies, equipment
(e.g., manual vs automated HLD), and drying/storage techniques. Moreover, study teams
varied at most of the sites. Therefore, reprocessing interventions varied at each
site but we attempted to standardize the approach to these interventions using published,
hierarchical Cascades and also statistically controlled for this variation using multivariable
analysis. It is also important to note that we used an established ATP threshold as
a benchmark to assess quality of manual cleaning, but cannot draw conclusions regarding
risk of infection transmission because we did not perform microbiological cultures
and ATP levels do not correlate well with culture data [9 ]. Moreover, the absence of a statistically significant difference in pre-and post-intervention
post-manual cleaning ATP in this study limits our conclusions about the impact of
interventions on subsequent phases of reprocessing. As noted previously, we also did
not assess whether additional interventions such as repeat manual cleaning or HLD
would further reduce post-manual cleaning ATP particularly in those endoscopes measuring
above the benchmark. Of note, we performed an air-purge of the instrument channel
to clear variable gross residual contents prior to sampling for ATP testing, which
despite being standardized, may have resulted in underestimation of bioburden. It
is also important to note that only a subset of all interventions made were directly
related to the manual cleaning phase and it is unclear how interventions made after
the post-cleaning phase impact post-cleaning ATP. We had planned to include two additional
training centers, but due to the COVID-19 pandemic, that was no longer possible and
thus, enrollment ended early. To our knowledge, ATP testing has not been well studied
in resource-limited settings, and we assumed that a cutoff of 200 RLUs should similarly
be applied as a benchmark of acceptable residual bioburden. Although we used locally
sourced tap water instead of sterile water at each site, ATP testing of tap water
revealed negligible results and there were no statistically significant differences
in post-manual cleaning ATP pre- and post-intervention. Finally, the baseline condition
of endoscopes between and within sites likely varied and may have introduced confounding
that we attempted to control for by retrieving endoscope service/repair history and
performing borescope assessment, but this was not be possible for every endoscope.
Conclusions
In conclusion, we found observation of reprocessing practices and ATP testing by an
experienced study team to be valuable in identifying opportunities to optimize reprocessing
at multiple training centers. Although this did not result in a statistically significant
reduction in post-manual cleaning ATP, the intervention was well received by training
center staff and adoptable. Further investigation is warranted to understand the significant
proportion of endoscope reprocessing cycles that fail to meet ATP benchmarks following
manual cleaning and how this can be addressed, especially in limited resource settings.
This might include study before and after endoscope channel and valve replacements,
aiming to remove potential sanctuaries harboring organisms despite standard endoscope
reprocessing.
