Introduction
Traditionally, there are only two methods for sedation in gastrointestinal endoscopy
(GIE): standard sedation (SS) and monitored anesthesia care (MAC). In the former,
intravenous drugs such as benzodiazepines and opioids are used, the target level is
moderate sedation and the method is supervised by an endoscopist. In MAC, the choice
of the drugs and the target level sedation are supervised by an anesthesiologist [1]
[2]. In recent years, in countries where non-anesthesiologist physicians can administer
propofol, a new method has emerged: non-anesthesiologist-administered propofol (NAAP).
In NAAP, propofol is administered alone or combined with other agents, the target
level sedation is moderate or deep [1]
[2]
[3], and an endoscopist with additional training in administration of sedo-analgesic
drugs and airway management supervises it. Thus, NAAP evolved from SS and is an alternative
to MAC.
Since the introduction of NAAP, we have collected a lot of information about its efficacy
and safety; currently, outcomes of endoscopist-directed sedation (EDS) are comparable
with SS and MAC [4]
[5]
[6]
[7]
[8]
[9]. In agreement with this data, a more recent study including 1.38 million procedures
with more than 880,000 colonoscopies and more than 508,000 gastroscopies compared
risk of serious adverse events with MAC and EDS. This study concluded that the safety
of the methods in colonoscopy was equivalent but that overall, MAC was riskier in
gastroscopy [4]. Similar data have been reported by other authors [5]
[6]
[7]
[8]. Furthermore, EDS during colonoscopy has been described as the safest among all
gastrointestinal endoscopic procedures performed with this sedation method [4]
[9]. Conversely, there is no information about the outcomes of colonoscopy itself when
the endoscopist has to fulfill the additional task of supervising the sedation. This
is probably due to the difficulty in measuring these outcomes because they depend
on various factors, such as equipment, endoscopist training, adequate bowel preparation,
and patient tolerance [10]
[11]
[12], among others. However, the advent of quality indicators for colonoscopy [13]
[14] has facilitated quantification of objective parameters and monitoring and comparison
of them.
The unquestionable advantages of sedation for GIE imply an additional task for the
endoscopist. At least in theory, this would distract the endoscopist and could have
a negative impact on colonoscopy. In this context, we have done this research to determine
whether the adenoma detection rate (ADR) in colorectal cancer (CRC) screening colonoscopies
performed with NAAP is equivalent to that for those performed with MAC)
Patients and methods
Study design
We conducted a single-blind, non-randomized controlled equivalence trial at a single
institution from January 2017 to December 2018. Colonoscopies were performed by two
expert endoscopists who had more than 7 years of experience in endoscopy and 1 year
of experience in colonoscopy for CRC screening. Both had ADRs > 50 % and acceptable
complication rates.
The two endoscopists alternated between colonoscopy schedules with NAAP and MAC.
In NAAP, an expert nursing team directed by an endoscopist administered the sedation.
This team had previously completed more than 3500 NAAP sedations for GIE. For MAC,
different anesthesiologists directed the sedation, according to their usual schedules.
After signing the consent form, patients were instructed to perform bowel preparation
following our usual colon cleansing protocol.
The local Research Committee and Research Ethics Committee approved our research (IRB
approval code: 42319_TDA_ANE. V2.0:02 /05 /19). The trial then was registered in ClinicalTrials.gov.
(NCT03922074). All authors reviewed and approved the final version of the manuscript.
Patients
Adults between 50 and 69 years old were recruited from the Spanish National CRC screening
program (CRCSP). After a positive fecal occult blood test (FOBT), they were blindly
assigned to undergo colonoscopy with either NAAP or MAC by the CRCSP office following
return of the FOBT results, patient availability for colonoscopy, and the availability
of slots in the endoscopy schedule. In our hospital, there are two schedules a week
for these colonoscopies: NAAP on Wednesday and MAC on Friday. The CRCSP secretary
did not know who administered sedation during each schedule.
Random assignment was not possible because no researcher took part in the allocation
of participants. Neither the anesthesiologists nor the endoscopists who directed sedation
allocated patients.
Patients with familial colorectal cancer history were excluded.
Intervention
In the NAAP arm, all patients were monitored throughout the procedure. We monitored
oxygen saturation, blood pressure, rhythm, and heart rate. Pre-oxygenation began 5
minutes before the beginning of intravenous (IV) drug administration. In all colonoscopies
with NAAP, a nurse exclusively administered the IV drugs and monitored the patient.
At the beginning of the procedures, we administered a fentanyl infusion in bolus:
75 μg (1.5 mL) and an initial propofol dose (10 mg/mL) in bolus of 0.5 to 2.5 mg/kg
followed by a maintenance dose of 20 to 60 mL/h through a target-controlled infusion
(TCI) pump. In this way, we achieved a moderate to deep level of sedation (Observer's
Assessment of Alertness/Sedation Scale [OAAS]: 1 to 3) [15].
For patients assigned to the MAC arm, the anesthesiologist directed the choice of
the IV drugs and the target level of sedation. The Anesthesiologist usually administered
propofol plus other agents (e. g. benzodiazepines, opioids, ketamine) in doses he
or she considered convenient.
Colonoscopies were performed per standard of care using Evis Exera III Video colonoscopes
CF-HQ190 and CF-H185 L/I by Olympus. Colonoscope insertion began with the patient
on the left lateral position, which was maintained until the cecum was reached. Auxiliary
maneuvers (e. g. specific abdominal pressure or change of the patient position) were
used as appropriate. Bowel preparation was evaluated using the Boston Bowel Preparation
Scale [16].
Colonic polyps were classified according to the Paris Classification [17] and their size, estimated by comparison of one with an open standard biopsy forceps.
Anatomic location and resection techniques also were documented. Finally, a case report
form was filled out for each participant immediately before and after the procedure.
Patients were followed up for 8 weeks by one researcher who reviewed their medical
records and documented polyp histopathology and the appearance of any adverse events
(AEs) associated with colonoscopy.
In addition, the cost of personnel for sedation was calculated by multiplying the
global exploration time by 0.89 euros when the anesthesiologist administered it and
0.49 euros when the nurse did it. These prices were the average personnel cost per
minute in our hospital.
Outcomes
Our main outcome measure was the ADR in CRCSP colonoscopies performed with NAAP. Secondary
outcomes measures were the advanced ADR (aADR), sessile serrated ADR (ssADR), mean
number of adenomas per procedure (MAP), and the complication rate (CR) associated
with these colonoscopies. We defined these indicators based on widely accepted definitions
[13]
[18]
[19]
[20]
[21]. Regarding sessile serrated adenomas (SSAs), we considered a histological definition,
i. e., a pathologist defined these lesions. In addition, SSAs ≥ 1 cm or SSAs with
dysplasia were considered advanced adenomas [18]
[19].
Sample size and statistical analyses
To assess the equivalence of ADR in CRC screening colonoscopies performed with NAAP
and performed with MAC, we assumed an expected ADR per group = 40 % [22], an equivalence margin = 10 %, a 95 % two-sided confidence interval (95 %CI), and
an allocation rate 1:1. With these assumptions Equivalence Tests for the Difference
between the two proportions were applied. Thus, 296 patients per group were calculated.
An initial descriptive analysis of the main variables was done. We estimated the measures
of central tendency and statistical dispersion. Chi-squared or Fisher exact tests
were used to determine whether there was a significant difference between the frequencies.
The t-test was used to compare the quantitative variables. The analyses were performed
on an intention-to-treat basis. Per-protocol analyses were used as sensitivity analyses
for the main outcome. There were no protocol violations.
Finally, the equivalence was confirmed when the 95 %CI for the difference between
both ADRs (NAAP vs. MAC) was entirely within the equivalence range of –0.10 to + 0.10.
Results
We included 630 patients with median age 59.76 ± 5.81 years, 40.5 % of whom were women
([Fig. 1]). The cecal intubation rate (CIR) was 97 %, adequate bowel preparation (ABp) was
81.8 %, the requirement for a withdrawal time > 6 minutes (WT6 m) was met in 98.7 %, and the global exploration time (ET) was 24.25 ± 8.86 minutes
(range, 8–70 min.). The ADR was 62.9 %, the advanced ADR (aADR) was 37.3 %, the sessile
serrated ADR (ssADR) was 5.2 %, and the mean number of adenomas per procedure (MAP)
was 1.53 ± 1.75. The complication rate (CR) was 0.6 %. All of the results except ET
were comparable in the MAC and NAAP groups, although that difference disappeared when
incomplete and inadequately prepared colonoscopies were excluded. The endoscopist
was the only factor associated with a significant difference in ITT and PP analysis
([Table 1] and [Table 2]).
Fig. 1 Flowchart showing the enrollment and the course of recruited patients during the
study. There were no losses during the follow-up. MAC, monitored anesthesia care;
NAAP, non-anesthesiologist-administered propofol; ITT, intention-to-treat; PP, per-protocol.
Table 1
Patient characteristics and quality indicators for and outcomes of colonoscopy: ITT
analysis.
|
Sedation type
|
MAC (n = 315)
|
NAAP (n = 315)
|
P value
|
|
Demographic data
|
|
|
56.83
|
62.22
|
0.168
|
|
|
59.37 ± 5.73
|
60.15 ± 5.87
|
0.093
|
|
Endoscopist
|
|
|
199 (63.17)
|
304 (96.51)
|
0.001
|
|
|
116 (36.83)
|
11 (3.49)
|
0.001
|
|
Quality indicators for colonoscopy
|
|
|
96.19
|
97.78
|
0.244
|
|
|
78.4
|
83.2
|
0.129
|
|
|
97.78
|
99.68
|
0.069
|
|
|
25.05 ± 8.4
|
23.44 ± 9.24
|
0.023
|
|
Outcomes of colonoscopy
|
|
|
61.59
|
64.13
|
0.564
|
|
|
40
|
34.6
|
0.187
|
|
|
5.41
|
5.08
|
0.860
|
|
|
1.42 ± 1.64
|
1.64 ± 1.85
|
0.110
|
|
|
0.63
|
0.63
|
1.000
|
ITT, intention to treat; MAC, monitored anesthesia care; NAAP, non-anesthesiologist-administered
propofol; ADR, adenoma detection rate; MAP, mean number of adenomas per procedure.
Table 2
Patient characteristics and quality indicators for and outcomes of colonoscopy: PP
analysis.
|
Sedation type
|
MAC (n = 247)
|
NAAP (n = 262)
|
P value
|
|
Demographic data
|
|
|
53.85
|
57.63
|
0.390
|
|
|
59.23 ± 5.69
|
60.15 ± 5.91
|
0.075
|
|
Endoscopist
|
|
|
158 (63.97)
|
253 (96.56)
|
0.001
|
|
|
89 (36.03)
|
9 (3.44)
|
0.001
|
|
Quality indicators for colonoscopy
|
|
Withdrawal time > 6 min. (%)
|
100
|
100
|
–
|
|
Exploration time (min.) (mean ± SD)
|
24.83 ± 7.75
|
23.42 ± 8.99
|
0.060
|
|
Outcomes of colonoscopy
|
|
|
61.94
|
62.98
|
0.810
|
|
|
38.87
|
33.21
|
0.196
|
|
|
4.07
|
4.58
|
0.830
|
|
|
1.40 ± 1.58
|
1.64 ± 1.91
|
0.137
|
|
|
0.81
|
0.76
|
1.000
|
PP, per-protocol; MAC, monitored anesthesia care; NAAP, non-anesthesiologist-administered
propofol; ADR, adenoma detection rate; MAP, mean number of adenomas per procedure.
Analysis by ITT showed an ADR in colonoscopies performed with MAC of 61.59 % compared
with 64.13 % performed with NAAP, difference (δADR): 2.54 %, 95 %CI: −0.10 to 0.05. Analysis
by PP showed an ADR in colonoscopies performed with MAC of 61.94 % compared with 62.98 %
performed with NAAP, δADR: 1.04 %, 95 %CI: −0.09 to 0.07. ([Table 3] and [Fig. 2]).
Table 3
Equivalence of ADR between MAC and NAAP.
|
MAC
|
NAAP
|
95 %CI δADR
|
|
ADR
|
|
|
61.59
|
64.13
|
–0.10–0.05
|
|
|
61.94
|
62.98
|
–0.09–0.07
|
|
|
|
|
ADR, adenoma detection rate; MAC, monitored anesthesia care; NAAP, non-anesthesiologist-administered
propofol; ITT, intention to treat; PP, per-protocol.
Fig. 2 95 % two-sided confidence interval for the difference between ADR (MAC vs NAAP). MAC,
monitored anesthesia care; NAAP, non-anesthesiologist-administered propofol; ADR,
adenoma detection rate; δADR, difference between ADR; ITT, intention-to-treat; PP,
per-protocol (analysis excluded incomplete and inadequately prepared colonoscopies).
One post-polypectomy bleed and one splenic injury occurred in the MAC group and one
post-polypectomy colonic perforation and one case of post-polypectomy syndrome occurred
in the NAAP group. No differences were found between the groups in CR: MAC 0.63 vs.
NAAP 0.63 (P = 1.0) ([Table 1]).
Regarding the endoscopists (E), E1 performed most of the included procedures, whereas E2 performed 36.83 % (n = 116) in the MAC group and 3.49 % (n = 11) in the NAAP group
(P = 0.001) ([Table 1]). E1 registered a superior CIR (98.41 % vs. 91.34 %, P = 0.01), more often met the WT6 m requirement (99.4 % vs. 96.06 %, P = 0.01) and achieved a higher ADR (64.61 % vs. 55.91 %, P = 0.08) and MAP (1.63 ± 1.81 vs. 1.15 ± 1.45, P = 0.01) ([Table 4]). When the incomplete and inadequately prepared colonoscopies were excluded, the
differences between ADR (64.48 % vs. 54.08 %, P = 0.06) and MAP (1.62 ± 1.83 vs. 1.12 ± 1.39, P = 0.01) were not only maintained but also increased, with a shorter ET (23.63 ± 8.07
vs. 26.09 ± 9.61 min., P = 0.01). However, in a subanalysis taking only Endoscopist 1’s results into account,
there were no differences between MAC and NAAP in any of the compared measures.
Table 4
Quality indicators for and outcomes of colonoscopy according to endoscopist (E).
|
E1 (n = 503)
|
E2 (n = 127)
|
P value
|
|
Demographic data
|
|
|
60.83
|
54.33
|
0.186
|
|
|
59.79 ± 5.83
|
59.65 ± 5.73
|
0.807
|
|
Quality indicators for colonoscopy
|
|
|
98.41
|
91.34
|
0.001
|
|
|
82.83
|
77.78
|
0.197
|
|
|
99.4
|
96.06
|
0.010
|
|
|
23.93 ± 8.48
|
25.5 ± 10.12
|
0.109
|
|
Outcomes of colonoscopy
|
|
|
64.61
|
55.91
|
0.081
|
|
|
36.58
|
40.16
|
0.473
|
|
|
5.78
|
3.15
|
0.274
|
|
|
1.63 ± 1.81
|
1.15 ± 1.45
|
0.002
|
|
|
0.6
|
0.79
|
1.000
|
ADR, adenoma detection rate.
The mean personnel cost per sedation during colonoscopy was significantly higher when
the anesthesiologist supervised the sedation: 22.29 ± 7.48 vs. 11.50 ± 4.50 euros,
P = 0.001. Thus, MAC was more expensive than NAAP: 7022.10 vs. 3623.55 euros, respectively.
Therefore, in our study, NAAP resulted in a savings of 3398.55 euros.
Discussion
Published reports to date indicate that NAAP is just as safe as MAC [4]
[5]
[6]
[7]
[8]
[9]. However, no information exists about NAAP’s impact on the most important aspect
endoscopy: procedure outcomes. Therefore, our research is the first to provide evidence
in favor of use of endoscopist-directed sedation for colonoscopy.
In our study, all quality indicators for colonoscopy except ABp, which has a minimum
standard of 90 %, were met [23]. This poor result was probably related to our colon cleansing protocol. Another
we recommended a split-dose bowel preparation, patients were not advised of when to
start and end it or to take the last dose as near as possible to their procedure time
for colonoscopies done in the morning and for afternoon colonoscopies, to take the
last dose on the same day [24]. Fortunately, those inaccuracies in the patient information have been corrected.
Regarding the outcomes of colonoscopy, our ADR largely exceeded the recommended cut-off
of 50 % for the CRC screening program in our region [25]. Conversely, a priori, our complication rate exceeded the accepted minimum standard
of 0.5 % [23]; however, it should be recognized that calculation of this rate only included our
sample in the denominator instead of all colonoscopies performed by the endoscopists
during the same period. In doing so, both endoscopists strictly met the requirement.
Furthermore, the AEs that occurred were following complex polypectomies, for which
a much higher complication rate is to be expected [26]
[27], and in our center, that rate was not exceeded [28].
Currently the ADR is the best quality indicator for colonoscopy and the only one strongly
related to interval CRC risk [13]
[29]; however, it is important to recognize that the ADR is an imperfect indicator [20]
[21]
[30]. For this reason, some reasonable alternative parameters have been proposed: advanced
ADR, sessile serrated ADR, and mean number of adenomas per procedure [20]
[21]
[30]
[31], among others. Until recently, however, no cut-off points had been established for
these measures [20]
[21]
[30], so we were not able to evaluate these results. In addition, a comparison of our
aADR, ssADR and MAP with other published reports would have been inaccurate given
the disparities among definitions, populations, and endoscopists. For instance, SSAs
have been defined based on histology in some studies or on their size and location
in other ones [31]
[32]
[33].
The demographics in our MAC and NAAP groups were comparable. In contrast, the bowel
preparation tended to be better and exploration time shorter in the NAAP arm, which
indicates that colonoscopies were performed more quickly in patients with adequate
preparation. Indeed, in this scenario, reaching the cecum is easier and less washing
is required; therefore, a thorough and complete mucosal inspection is accomplished
in a shorter time [23]. On the other hand, as discussed below, our findings would have been influenced
by better performance by Endoscopist 1, who performed more procedures in the NAAP
arm. When we excluded the incomplete and inadequately prepared colonoscopies, the
differences between NAAP and MAC disappeared.
Regarding the outcomes of colonoscopy, there were no differences between MAC and NAAP
in ADR, complication rates, or other calculated measures, even when the incomplete
and inadequately prepared colonoscopies were excluded.
In our study the equivalence margin, to assess the equivalence of ADR in CRC screening
colonoscopies performed with NAAP and performed with MAC, was 10 % so the bounds of
the confidence interval were –0.10 and + 0.10. In PP analysis, the 95 %CI of the difference
between both ADR lay entirely within the equivalence range, Consequently, this trial
has confirmed that the ADR with NAAP is equivalent to the ADR with MAC. Nevertheless,
it should be noted that our results reflect a comparison between an anesthesiologist
and an expert nursing team directed by an endoscopist in which the hypothetical additional
distraction and difficulty for the endoscopist, if it existed at the beginning, had
already been overcome.
Concerning the two endoscopists (E) who took part in the study, there was a clear
trend toward a higher ADR in favor of E1 and a significant difference between MAP and NAAP for that endoscopist. The ADR reflects
adequate inspection of the bowel mucosa [23], which in turn depends on, among other things, a complete and thorough evaluation.
CIR and WT6 m, respectively, directly determine these two last requirements [23], and those were significantly higher for E1 as well. An upward trend in ADR is very
important because it has been proven that, in a primary colonoscopy screening setting,
a 1 % increase in ADR predicted a 3 % decrease in the risk of interval CRC [29]. On the other hand, although there is no accepted cut-off for MAP, we agree that
in the proper circumstances, use of this indicator would be a complementary way to
evaluate and compare endoscopist performance [21]. The superior results by E1 may have been associated with his training and longer
experience with CRC screening colonoscopy, which has been demonstrated to improving
endoscopist skills [12]. Despite the clear differences between the endoscopists and their contributions
to the study, the ADRs for MAC and NAAP were probably equivalent because the sedation
method did not affect endoscopist performance, as was indicated by the subanalysis
of Endoscopist 1’s results.
Finally, even though this trial was not an economic study, our personnel costs indicate
that NAAP results in a concrete and significant savings, which would be even more
significant if use of this method of sedation was expanded to other health systems.
Conclusions
In summary, from the previous data, we can conclude that ADR in colorectal cancer
screening colonoscopies performed with NAAP is equivalent to ADR in colonoscopies
performed with MAC. Similarly, there is no difference in complication rates.
In spite of the obvious limitations of our study, such as the absence of randomization
or the fact that it was conducted at a single institution, we were able to propose
an economical, effective, safer and more accessible alternative to traditional MAC
without decreasing outcomes of colonoscopy. The result may be a solution to the increasing
problem of lack of expert personnel in most national health systems, and specifically
for supporting widespread CRC screening programs.
