Introduction
About 17 to 20 million endoscopy procedures including esophagogastroduodenoscopy (EGD),
flexible sigmoidoscopy and colonoscopy are performed in the United States annually
[1]
[2]. Majority of these procedures require some form of sedation to reduce patient anxiety,
minimize patient discomfort and maximize patient safety thus allowing the endoscopist
to perform a thorough examination [3]. Traditionally, moderate sedation has been used to perform these procedures. It
is a pharmacologic induced state in that the patient is still conscious but is able
to tolerate unpleasant stimuli and maintain cardiovascular status [4]. Even though moderate sedation is safe, it is associated with minor or major adverse
events (AEs) in 1 per 170 to 1 per 10,000 endoscopy procedures and cardiorespiratory
events make up for majority of these AEs [5]. Therefore, patients are continuously monitored during moderate sedation to avoid
excessive sedation and cardiorespiratory depression. Obesity is one of the factors
that make administration of moderate sedation challenging. It can affect the metabolism
of most drugs used in moderate sedation altering the dose required to achieve effective
sedation [6]. Obesity is typically associated with various cardiopulmonary comorbidities that
may increase the rate of cardiorespiratory AEs during the procedure [7]
[8]
[9]
[10]. However, to date there is very limited literature on the outcomes of obese patients
undergoing endoscopy procedures with moderate sedation. The aim of this study was
to evaluate the outcomes of general endoscopy procedures performed with moderate sedation
in patients with obesity.
Patients and methods
This was an Institutional Review Board (IRB) approved retrospective study that included
all patients undergoing EGD or colonoscopy at University of Arkansas for Medical Sciences
(UAMS) from July 1, 2017 to December 31, 2019. All adults (> 18 years of age) who
had EGD and/ or colonoscopy with moderate sedation at UAMS during the study period
were included. Patients who had: 1) an EGD or colonoscopy performed with the involvement
of an anesthesia provider; 2) EGD or colonoscopy without sedation; 3) procedures other
than EGD or colonoscopy; 4) colonoscopies with documented poor bowel preparation (Supplementary file 1); and 5) colonoscopies that were performed together with an EGD were excluded from
the study. The rationale for last criteria was that these patients had already had
sedation for the EGD and therefore, we were not able to correctly assess the sedative
medications administered for these colonoscopies.
End points for patients undergoing EGD were 1) total procedure time; 2) recovery time,
3) immediate cardiopulmonary AE; 4) 7-day all-cause hospitalization after the procedure,
5) poor tolerance of moderate sedation; 6) doses of medications used for the procedure;
and 7) use of adjunctive sedatives for sedation during the procedure. For patients
undergoing colonoscopies cecal intubation time (CIT) and withdrawal time were considered
as end points in lieu of total procedure time.
Moderate sedation was defined as sedation directed by the physician performing the
endoscopic procedure and provided by a nurse during the procedure using Midazolam
and Fentanyl. Poor bowel preparation was defined as clear documentation of bowel prep
as being “poor” or rated as < 6 on Boston Bowel Prep Score. Adjunctive sedatives were
medications that are not primarily used for sedation but have strong sedative effect
and are used in conjunction with Midazolam and/ or Fentanyl i. e. diphenhydramine
and prochlorperazine. Total procedure time was the time from scope insertion to complete
withdrawal of the scope. Cecal intubation time was time from insertion of the colonoscope
to the time cecum was reached. Withdrawal time for colonoscopies was the time from
cecal intubation time to complete withdrawal of the colonoscope. Recovery time was
the time patient spent in the postoperative section of the endoscopy unit before being
discharged or sent back to floor. Immediate cardiopulmonary AEs were cardiac or pulmonary
AEs that were noted during or immediately after the procedure and were documented
in the chart including hypotension and hypoxia. Poor tolerance of moderate sedation
was assessed based on the documentation in the endoscopy report. This documentation
included clear documentation of tolerance of procedure, reordering the procedure with
anesthesia or requirement of anesthesia for future procedures. Repeat procedures with
anesthesia for therapeutic purposes was not counted as poor tolerance of the procedure.
Patients undergoing EGD or colonoscopy procedures were identified with the most commonly
used administrative codes associated with both types of procedures (Supplementary file 1). Body mass index (BMI) recorded at the time of the procedure. Information about
demographics (age, gender, race), history of smoking or alcohol use and provider performing
the procedure (gastroenterologist, hepatologist, other) was obtained on each patient.
Other provider category included surgeons and family medicine physicians. These providers
had performed very few (15) EGDs in total for non-obese and obese patents. Therefore,
these procedures were excluded from the analysis. All the providers had been in full
time practice for ≥ 2 years and performed ≥ 250 colonoscopies per year. Information
on all primary and secondary end points was collected. Co-morbid illness assessment
was done by calculating the Elixhauser Comorbidity Index (ECI) score using the Van
Walraven algorithm [11]
[12]
[13]
[14]. Obstructive sleep apnea (OSA) was recorded as a separate comorbidity because it
is not included in ECI and is an important factor to consider during sedation. Number
of prescription medications listed in each patient’s medication list that can affect
tolerance or efficacy of moderate sedation were also recorded (Supplementary file 1). Patients were divided into three groups: 0, 1 or ≥ 2 prescription medications with
sedative effects. Details of pre-procedure assessment and administration of moderate
sedation are provided in
Supplementary file 1.
Descriptive statistics were used to perform exploratory analyses. Categorical data
were described as proportions and analyzed using chi-square test or Fisher exact test.
Continuous data were described as mean (standard deviation or SD) or median (interquartile
range [IQR]) and analyzed using t-test, ANOVA, Wilcoxon rank-sum test or Kruskal-Wallis test depending on the distribution
of the variable. Data were analyzed separately for EGDs and colonoscopies performed
with moderate sedation. Some patients underwent more than one EGD or colonoscopy.
Therefore, per-procedure analysis was conducted. Per-patient sensitivity analysis
was conducted to determine if repeat procedures for the same patient were affecting
the study results. Exploratory analysis using standard statistical methods showed
that the obese and non-obese groups were dissimilar to each other age, gender, race,
number of prescribed medications with sedative effects, Elixhauser comorbidity index
score, OSA and tobacco and alcohol use. Therefore, propensity score matching (PSM)
was performed to make the study groups comparable to each other. Data were stratified
by EGD and colonoscopy and 1:1 match was conducted with propensity score margin being
set at 0.1. Logistic regression was performed separately for EGDs and colonoscopies
to identify procedure related factors associated with obesity. Subgroup analysis was
conducted by obesity class for EGDs and colonoscopies separately. The methodology
and results of this analysis is provided in the Supplement file. All OR values were
reported up to two decimal points except for values that would have rounded to 1.00.
Two-sided P < 0.05 was considered significant. The analysis was performed with SAS software version
9.4 (SAS Institute Inc., North Carolina, United States). Additional details of statistical
analysis can be found in the supplement file.
Results
Data description and PSM
A total of 7758 procedures were performed with moderate sedation during the study
period. One or more variables were missing for one hundred and fifty-seven (2.02 %)
procedures and these were excluded. Final analysis included 7601 procedures performed
with moderate sedation for 5746 patients. Out of these 3545 (46.64 %) were EGDs and
4056 (53.36 %) were colonoscopies. Propensity score matching identified 1360 and 1740
pairs of EGDs and colonoscopies with moderate sedation for patients with and without
obesity. [Table 1] summarizes the demographic and clinical variables and [Table 2] summarizes the procedure related outcomes of patients undergoing EGD and colonoscopy.
Table 1
Demographics and clinical characteristics of patients with and without obesity undergoing
EGD and colonoscopy with moderate sedation after propensity score matching.
|
EGD
|
Colonoscopy
|
|
Non-obese
(BMI < 30;
n = 1360)
|
Obese
(BMI ≥ 30;
n = 1360)
|
P value
|
Non-obese
(BMI < 30;
n = 1740)
|
Obese
(BMI ≥ 30;
n = 1740)
|
P value
|
|
Mean age in years (SD)
|
57.12 (16.02)
|
55.68 (13.47)
|
0.01
|
57.46 (13.74)
|
57.21 (12.04)
|
0.56
|
|
Female (%)
|
839 (61.69)
|
886 (65.15)
|
0.06
|
1049 (60.29)
|
1161 (66.72)
|
< 0.0001
|
|
Race (%)
|
|
|
0.88
|
|
|
< 0.001
|
|
|
936 (68.82)
|
929 (68.31)
|
1104 (63.45)
|
977 (56.15)
|
|
|
368 (27.06)
|
370 (27.21)
|
558 (32.07)
|
713 (40.98)
|
|
|
56 (4.12)
|
61 (4.49)
|
78 (4.48)
|
50 (2.87)
|
|
Prescribed medications with sedative effects (%)
|
|
|
0.06
|
|
|
< 0.01
|
|
|
788 (57.94)
|
820 (60.29)
|
1001 (57.53)
|
907 (52.12)
|
|
|
301 (22.13)
|
252 (18.53)
|
384 (22.07)
|
419 (24.08)
|
|
|
271 (19.93)
|
288 (21.18)
|
355 (20.40)
|
414 (23.79)
|
|
Median Elixhauser comorbidity index (IQR)
|
4 (0–10)
|
5 (-4–9)
|
< 0.0001
|
0 (0–5)
|
-2 (-4–5)
|
< 0.0001
|
|
Obstructive sleep apnea (%)
|
57 (4.19)
|
185 (13.6)
|
< 0.0001
|
90 (5.17)
|
304 (17.47)
|
< 0.0001
|
|
Tobacco use (%)
|
208 (15.29)
|
169 (12.43)
|
0.03
|
293 (16.84)
|
299 (17.18)
|
0.79
|
|
Alcohol use (%)
|
114 (8.38)
|
123 (9.04)
|
0.54
|
76 (4.37)
|
57 (3.28)
|
0.09
|
|
Procedure done as outpatient (%)
|
1017 (74.78)
|
1083 (79.63)
|
< 0.01
|
1591 (91.44)
|
1604 (92.18)
|
0.42
|
|
Endoscopist performing the procedure
|
|
|
< 0.001
|
|
|
< 0.0001
|
|
Gastroenterologists
|
857 (63.01)
|
762 (56.03)
|
892 (51.26)
|
890 (51.15)
|
|
Hepatologists
|
503 (36.99)
|
598 (43.97)
|
551 (31.67)
|
635 (36.49)
|
|
Other providers
|
--
|
--
|
297 (17.07)
|
215 (12.36)
|
BMI, body mass index; EGD, esophagogastroduodenoscopy; IQR, interquartile range;
SD, standard deviation.
Table 2
Procedure-related outcomes of patients with and without obesity undergoing EGD and
colonoscopy with moderate sedation after propensity score matching.
|
EGD
|
Colonoscopy
|
|
Non-obese (BMI < 30; n = 1360)
|
Obese (BMI ≥ 30; n = 1,360)
|
P value
|
Non-obese (BMI < 30; n = 1740)
|
Obese (BMI ≥ 30; n = 1,740)
|
P value
|
|
Median procedure time in minutes (IQR)
|
6 (4–9)
|
5 (4–8)
|
< 0.01
|
21 (16–28)
|
21 (16–29)
|
0.75
|
|
Median Cecal intubation time in minutes (IQR)
|
--
|
--
|
--
|
7 (5–11)
|
7 (5–11)
|
< 0.01
|
|
Median withdrawal time in minutes (IQR)
|
--
|
--
|
--
|
12 (9–18)
|
13 (9–18)
|
0.41
|
|
Median recovery time in minutes (IQR)
|
32 (30–39)
|
32 (30–37)
|
0.41
|
31 (30–36)
|
31 (30–36)
|
0.79
|
|
Intra-procedure cardiopulmonary adverse events (%)
|
15
(1.10)
|
10
(0.74)
|
0.32
|
7
(0.40)
|
7
(0.40)
|
1.00
|
|
7-day all cause hospitalization (%)
|
27
(1.99)
|
23
(1.69)
|
0.57
|
33
(1.90)
|
28
(1.61)
|
0.52
|
|
Poor tolerance of procedure or aborted procedure (%)
|
51
(3.75)
|
42
(3.09)
|
0.34
|
86
(4.94)
|
75
(4.31)
|
0.37
|
|
Median dose of midazolam used in mg (IQR)
|
5 (4–7)
|
5 (4–7)
|
< 0.01
|
5 (4–7)
|
5 (4–7)
|
0.24
|
|
Median dose of fentanyl used in mcg (IQR)
|
75 (50–100)
|
100 (50–100)
|
0.01
|
100 (75–125)
|
100 (75–125)
|
0.86
|
|
Use of adjunctive sedation medications (%)
|
191
(14.04)
|
155
(11.4)
|
0.04
|
174
(10.00)
|
167
(9.60)
|
0.69
|
BMI, body mass index; EGD, esophagogastroduodenoscopy; IQR, interquartile range.
Comparison of obese and non-obese patients after PSM
Median procedure time was shorter for obese patients (5 minutes, IQR 4 vs. 6 mins,
IQR 5; P < 0.01). For colonoscopies, cecal intubation time was slightly longer in obese patients
(7 minutes, IQR 6 vs. 7 minutes, IQR 6; P < 0.01; [Table 2]). For EGDs, median administered dose of Midazolam (5 mg, IQR 3 vs. 5 mg, IQR 3;
P < 0.01) and fentanyl (100 mcg, IQR 50 vs. 75 mcg, IQR 50; P = 0.01) were higher for obese patients ([Table 2]). Adjunctive sedatives were used more commonly in non-obese patients undergoing
EGDs (14.04 % vs. 11.4 %; P = 0.04; [Table 2]). Other procedure-related outcomes for EGDs and colonoscopies were similar between
the two groups ([Table 2]).
Regression model
Recovery time was found to be shorter for obese patients undergoing EGD (OR: 0.989,
95 % CI: 0.981–0.998; P = 0.01). None of the other procedure related outcomes were different between patients
with and without obesity for EGDs or colonoscopies.
Per-patient sensitivity analysis
Baseline population characteristics and regression model results from per-patient
analysis for both EGDs and colonoscopies were similar to per-procedure analysis.
Discussion
This is the first study to evaluate procedure-related outcomes in obese patients undergoing
endoscopy (EGD or colonoscopy) with moderate sedation. All procedure related outcomes
for EGDs and colonoscopies performed with moderate sedation were similar between patients
with and without obesity except for a shorter recovery time for obese patients undergoing
EGD (OR: 0.99, [Table 2]). However, the magnitude of this difference was so small that OR had to be reported
to 3 decimal points and the median recovery time for EGDs for both obese and non-obese
patients was the same (32 minutes, [Table 2]). This means that the statistical difference most likely happened from observations
at the extremes (< 25th or > 75th percentile) that happen far less frequently than
observations clustered around the center (between 25th and 75th percentile). Our results
are consistent with a prior study that reported no association of obesity with increase
in procedure time and administered doses of fentanyl or midazolam [15].
Data on use of moderate sedation and their outcomes in obese patients undergoing endoscopy
is limited to only a few studies that have focused on intra-procedure hypoxia or post-procedure
patient comfort [16]
[17]
[18]. Obesity is known to alter the upper airway anatomy. Increased fat tissue in the
head and neck areas predispose these patients to increased risk of obstructive sleep
apnea (OSA), application of non-invasive and invasive ventilation devices during sedation
[19]
[20]. This is compounded by a higher risk of impaired cardiopulmonary function and altered
metabolism of sedative medications in obese patients [6]
[7]
[8]
[9]
[10]. OSA was found to be significantly more prevalent in patients with obesity in our
population as well ([Table 1, ]
[Table 3]). All of these factors raise the theoretical concern of failure of moderate sedation
or increased AEs and recovery time from moderate sedation in these patients. However,
in our study no difference in sedative requirements was noted in obese patients compared
to non-obese individuals. This was noted even for adjunct sedatives. Similarly, cardiopulmonary
AEs and 7-day post-procedure hospitalizations were similar between obese and non-obese
patients. Overall, our study suggests that moderate sedation can be used safely and
effectively for performing endoscopic procedures in obese patients including patients
with OSA, which is consistent with previous literature [21]. This is a critical finding especially because obesity rate, endoscopy procedure
volumes, its associated costs, and use of anesthesia for endoscopy procedures are
all on the rise in the United States [1]
[2]
[22]
[23]
[24]. Judicious use of moderate sedation for certain endoscopic procedures in obese patients
can help offset the burden on limited anesthesia resources and associated costs.
Table 3
Logistic regression model showing procedure related factors associated with obesity
in patients undergoing EGD and colonoscopy with moderate sedation
|
OR (95 % CI)
|
P value
|
|
EGD
|
|
|
0.989 (0.981-.998)
|
0.01
|
EGD, esophagogastroduodenoscopy; OR, odds ratio.
Unsedated procedures are an alternative option in patients deemed at high risk of
sedation-related AEs. Two separate studies showed high success rate (92–98 %) of transnasal
EGDs without sedation using an ultrathin scope in patients with morbid obesity (BMI > 40 kg/
m2) [25]
[26]. The sample size in both studies was small (25 and 98), study populations were heterogenous
(Veteran Affairs patients and patients in Brazil) and the EGDs were specifically done
as a part of pre-bariatric surgery evaluation. Although, outcomes of unsedated colonoscopies
have not been specifically studied in patients with obesity, colonoscopy without sedation
has been performed with high success rate and patient satisfaction in selected patient
groups [27]
[28]. Therefore, unsedated procedures can be considered in selected patients with obesity
after a detailed risk and benefit discussion. Further studies are needed to specifically
assess tolerance and satisfaction of unsedated procedures in patients with obesity.
This study has several strengths. The sample size was large (n = 7601) and complete
data were obtained for 98 % of procedures with only 2 % of procedures having missing
data. Propensity score matching was conducted to make the obese and non-obese populations
as similar to each other as possible. A large number of variables were studied including
multiple procedure-related outcomes (7 for EGDs and 8 for colonoscopies) and patient
related factors (9 each for EGD and colonoscopy including prescription medications
that can alter the effect of moderate sedation). None of the prior studies have looked
at as many outcomes and patient variables. Additional analysis was performed for obesity
classes to study the effect of increasing obesity on procedure related outcomes. Lastly,
the AUC for regression models for EGD and colonoscopy were 0.63 and 0.69, respectively.
This is indicative of the robustness and reliability of the data analysis. In other
words, the regression models accounted for 63 % to 69 % of the total variability in
the data that can only be obtained with a large sample size and after accounting for
major counfounders [29].
This study had some limitations as well. It was retrospective in nature that is associated
with the inherent limitations of such studies. Reporting of patient intolerance in
the procedure report by the endoscopist may have been limited. The intent of this
study was to primarily to look at the type of intolerance that was severe enough to
either abort the procedure, cause a cardiopulmonary AE or significant discomfort to
the patient that necessitates use of anesthesia in the future. We were able to obtain
information on that type of intolerance from the procedure reports. There was also
a possibility of selection bias where endoscopic procedures for obese and morbidly
obese patients with several comorbidities were performed with anesthesia and obese
patients with fewer comorbidities had their procedures with moderate sedation. This
limitation was addressed by performing PSM. Although PSM is not the same as randomization
in a randomized controlled trial, this was the best statistical method available to
control for known confounders. Finally, the diagnosis of OSA was based on chart review
and could have been underdiagnosed in the study population. However, it was still
found to be an independent predictor in the logistic regression models for both EGD
and colonoscopy.
Conclusions
In conclusion, the results of this study strongly suggest that moderate sedation is
well tolerated by obese patients undergoing EGD or colonoscopy without a clinically
significant difference in most procedure related outcomes, AEs or post-procedure hospitalization.
Further studies are needed to confirm these findings.
