Aims To measure and compare, in a real-world setting, the frequency of adverse events
for patients undergoing procedural sedation and analgesia (PSA) during endoscopic
procedures between standard of care monitoring (SOCM) and the additional use of capnography.
Capnography measures exhaled carbon dioxide, providing real-time identification of
airway obstruction and respiratory depression.
Methods A before-and-after, quality-improvement protocol was implemented at five centres
in Leuven, Toronto, Ankara, Madrid, and Cambridge. Adverse events collected were escalation
of care, mild (75–90%; up to 60 seconds (s)) and severe (< 75% or<90% for over 60
s) oxygen desaturation, prolonged apnoea (> 60 s), bradycardia, tachycardia, airway
obstruction, cardiovascular shock or collapse, cardiac arrest, and death. Data were
collected between 2017 and 2021 per procedure using SOCM or SOCM plus capnography
(capnography). Patients could have more than one procedure performed. Data collection
started without modification to each centre’s standard practice for PSA. SOCM during
PSA included electrocardiogram, respiratory rate, blood pressure, and pulse oximetry.
Capnography was then introduced, and staff were trained on its use. Following a 2–4-week
wash-in period, data collection continued for the capnography cohort.
Results Data were collected for 6,801 procedures (n=3,431 for SOCM and n=3,370 for capnography).
Average procedure times were 25 minutes without significant differences between both
arms. Procedures were split into colonoscopy (n=2,773), gastroscopy (n=1,868), combined
gastroscopy and colonoscopy (n=502), endoscopic ultrasound (n=676), bronchoscopy (n=449),
endoscopic retrograde cholangiopancreatography (n=452), and others (n=81). When comparing
SOCM versus capnography, the relative risk (RR) of at least one adverse event occurring
per procedure was 0.66 (95% confidence interval: 0.56–0.78; 11.25% vs 7.19%). Escalation
of care decreased with capnography RR 0.08 (0.01–0.65; 12 vs 1 events). Oxygen desaturation
also decreased by RR 0.66 (0.54–0.80; 243 vs 158 events) for mild and RR 0.55 (0.32–0.96;
35 vs 19 events) for severe cases. Bradycardia and tachycardia were reduced using
capnography by RR 0.42 (0.24–0.74; 41 vs 17 events) and RR 0.15 (0.07–0.33; 48 vs
7 events), respectively. Prolonged apnoea was increased by RR 4.33 (1.46–12.85; 4
vs 17 events) when using capnography vs SOCM, likely due to an easier detection of
apnoea when using capnography. Differences in airway obstruction and cardiovascular
events were not significant and no patients died.
Conclusions A significant decrease in adverse events occurring during PSA were observed when
using capnography monitoring in addition to SOCM versus SOCM only in this initial
analysis of a large, multi-centre, quality-improvement initiative in five countries.
Additional analyses are required to limit the effect of confounders on results.
