Keywords Quality and logistical aspects - Hygiene - Endoscopy Upper GI Tract - Quality management
Introduction
Digestive endoscopy has been proven to produce aerosols [1 ]
[2 ]
[3 ]. This represents a risk of infection by COVID-19 and other airborne viruses. The
World
Health Organization (WHO) has defined aerosols as particles < 5 μm in diameter, which
can
remain airborne for many hours and can deposit in the lower airways to cause infection,
and
droplets as particles 5 to 10 μm in diameter, which quickly settle due to gravity
but may
contaminate surfaces [4 ]
[5 ]. A number of protective barriers have been proposed to minimize that risk. Continuous
suction of the oral cavity [1 ], shielding barriers [7 ]
[8 ], masks [9 ]
[10 ], and increasing the distance between patient and endoscopist [11 ] and improved ventilation such as laminar flow theaters [12 ] have been proposed as methods to reduce the exposure of endoscopists and endoscopy
staff to aerosols. Here, we present a study that uses masks that are clinically approved
for
bronchoscopy (Explorer endoscopy facemask, Intersurgical Ltd., United Kingdom) to
attenuate
aerosol production at the patient’s mouth (bare mask shown in [Fig. 1 ]
a and in use during an upper gastrointestinal endoscopy in
[Fig. 1 ]
b ). We find that this approach offers 47% (P = 0.01) reduction in particle count for particles < 5 μm in diameter (i.e.
aerosols), which are known to spread SARS-CoV-2.
Fig. 1 Effect of mitigations on aerosol count. a Photograph from procedure showing application of bronchoscopy mask to patient. b Effect of bronchoscopy masks, showing significant reduction when being used in the
< 5-µm diameter range. *P < 0.05, **P < 0.01.
Materials and methods
To establish the effect of the masks, we measured 12 upper gastrointestinal endoscopy
procedures in which the masks were placed on patients immediately before administering
xylocaine anesthetic throat spray and were removed after final oral extubation. As
a control we measured 37 procedures using normal clinical protocols. This was a prospective
study: We randomly selected patients to wear or not wear the protective mask and the
outcome of interest was the amount of aerosol produced from the patient’s mouth. A
priori power calculations based on limited previous studies determined that with five
replicates per patient, we can detect an effect size (Cohen’s d) of 1.98, sufficient
to differentiate between a cough and sneeze [13 ]
[14 ]. We also conducted a retrospective power analysis to indicate sample sizes that
may be required for further studies.
For both patient groups, procedures were performed across several endoscopy rooms
within the same endoscopy suite (Treatment Centre, Nottingham University Hospitals
NHS Trust), all of which had similar room ventilation at 15 to 17 air changes per
hour (measured using a balometer), similar size, air temperature, and humidity levels.
Particle counts were measured and analyzed using an AeroTrak portable particle counter
(TSI, Shoreview Minnesota, United States, model 9500–01) with inlet tube placed 10
cm from the patient’s mouth (methodology described in [3 ]) to maximize detection sensitivity and for compatibility with previous studies [1 ]. This device measures particles in six diameter ranges (0.5–0.7 μm, 0.7–1.0 μm,
1.0–3.0 μm, 3.0–5.0 μm, 5.0–10.0 μm, 10.0–25 μm), but for simplicity of analysis,
we grouped particles into aerosol and droplet size ranges as defined by the WHO and
used in previous studies [3 ]
[5 ]. For greater detail about spatial dispersal around the room, several particle counters
could be run simultaneously in different locations as previous studies have done [15 ], although care must be taken to avoid reduction in instrument sensitivity because
distance from the particle source is increase [16 ]. All personnel present in the room wore enhanced personal protective equipment (PPE),
which minimized the contribution of additional human aerosol sources. Staff and patients
were asked to remain as still as possible during recording to avoid scattering of
dust or re-aerosolization of liquids on surfaces that might cause increased particle
counts. Any unavoidable major movements of people were recorded and time-stamped so
they could be excluded if necessary. Previous studies have shown that when particles
are measured in this way, there is no significant contribution from events such as
biopsy, insertion/removal of catheters, insufflation or diathermy cutting [3 ]. Therefore, we do not expect that such events would have significantly impacted
the results here.
We compared aerosol and droplet concentrations produced from whole procedures (median
duration of 7.2 minutes), but we normalized counts to a 20-minute procedure by multiplying
total particle count by the appropriate ratio. All statistical analysis was performed
using the MATLAB software package (The MathWorks Inc., Massachusetts, United States).
Building on existing models of aerosol production in the respiratory tract, we used
a log-normal distribution to model the distribution of total particle counts [17 ]. For the whole-procedure data, a logarithm of the data was first computed, then
a t-test was applied to compute P values. For individual events, we compared particle counts in 30-second windows before
and after an annotated event (e.g. intubation) as per our previously published methodology
[3 ]. For events, the data distribution was modeled as the sum of a log-normal and normal
distribution to account for negative values of particle counts that can arise from
the subtraction step. A Monte-Carlo sampling method, therefore, was used to provide
numerical estimates of P value and numerically estimate mean ratios and confidence intervals between events
[18 ].
Results
Health Research Authority and ethical approval was granted by the Wales Research Ethics
Committee prior to the start of the study (IRAS no. 285595). We included patients
undergoing routine upper gastrointestinal endoscopy on the lists of 13 different participating
endoscopists at the Endoscopy Unit of the Nottingham University Hospitals NHS Trust
Treatment Centre between October 2020 and March 2021. The inclusion criteria were
adult patients > 18 years with capacity to consent. Biographical data from the patients
is shown in [Table 1 ]. We found that over the period when the bronchoscopy mask was attached, the total
number of aerosol-sized particles produced was reduced by 47% (95% CI: 16.8%-65.6%,
P < 0.01) compared to without masks [Fig. 1 ]
c . We did not find a significant reduction in total particle count for the droplet
range (≥ 5 µm). Considering individual events, we found that the key aerosol-generating
events of coughing, extubation, and anesthetic throat spray application were not significantly
reduced when the masks were used.
Table 1 Summary table showing demographic data for patients enrolled in this study.
Mask scenario
variable
No mask on patient
Bronchoscopy mask on patient
BMI, body mass index.
n
37
12
Age
Range: 24–93
Median: 61
Range: 41–83
Median: 75.5
Sex
Male: 23, Female: 14
Range: 17-38
Median: 26.9
BMI
Range: 16.3–38.2
Median: 24.8
Range: 17–38
Median: 26.9
Smoking
Smoker: 10
Non-smoker: 27
Smoker: 1
Non-smoker: 10
Vaper: 1
Hiatal hernia
Yes: 10, No: 27
Yes: 5, No: 7
Sedation
Midazolam: 16
Throat spray only: 21
Midazolam: 5
Throat spray only: 7
Discussion
Based on our whole-procedure analysis, we recommend that bronchoscopy masks or similar
be used to mitigate aerosols during outbreaks of respiratory diseases such as COVID-19.
However, our analysis of individual events suggests that although the masks are effective
at containing continuous low-volume aerosols production, e.g. breathing, they are
less effective at containing fast, high-volume production events. We suggest that
this is due to the openings in the mask required for breathing and the relatively
constant rate of suctioning: If aerosol production events exceed the suction rate,
these particles will necessarily escape via these holes. Further reduction for individual
events may require the use of negative pressure masks [18].
There are a number of limitations and remaining questions following this study. First,
particles greater than 5 µm in diameter (droplets) do not appear to be greatly impacted
by the masks. This may because they are relatively low in number and that a larger
sample size is needed to see this effect. Our retrospective analysis of study power
suggests that given the measured data in the aerosol size range with our sample size
of n = 12, study power is 0.88, which is comfortably above the threshold of 0.8 usually
required for such studies. However, larger studies in the future may be required to
better predict, understand, and mitigate flow, distribution, and elimination behaviors
of aerosols and droplets. Further, it will also be important to measure a wider range
of procedures and to use particle counters in different room positions. This may help
to examine the impact of other sources of aerosols in the rooms and to examine the
impact of events such as biopsies, diathermy cutting, and insertion/removal of catheters,
although our previous work did not find these to be significant producers of particle
when measured near a patient’s mouth [3 ]. Finally, although the endoscopists did not note significant reduction in maneuverability
of the endoscope due to the presence of the mask, future studies should evaluate this
thoroughly by performing post-procedure surveys of different endoscopists.
Conclusions
Overall, the reduction in particle levels may be sufficient to warrant reduced fallow
time because fewer particles mean shorter air clearance time, but is not sufficient
to eliminate the need for PPE for healthcare staff. We recommend that improved masks
be designed that can mitigate aerosols more effectively.
