Keywords
CRC screening - Endoscopy Lower GI Tract - Polyps / adenomas / ... - Endoscopic resection
(polypectomy, EMR, ...)
Introduction
Meta-analyses [1]
[2], randomized controlled trials [3]
[4]
[5], and observational studies [6], have demonstrated that prophylactic placement of hemostatic clips is effective
in preventing delayed hemorrhage for non-pedunculated colorectal lesions ≥ 20 mm in
size, located proximal to the splenic flexure, and removed using electrocautery. All
three criteria must be met before colorectal lesions have a sufficient risk of delayed
hemorrhage that allows prophylactic clip closure to be demonstrated effective from
both clinical and cost perspectives [1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]. For lesions that do not meet these three criteria, there is no clear evidence that
prophylactic clip placement is effective in preventing delayed hemorrhage.
Colorectal lesions can also be removed without electrocautery. This process is termed
cold resection and can be performed with or without submucosal injection. Cold resection
using snares has been shown to be effective for all colorectal lesions < 10 mm in
size [9], as well as sessile serrated lesions of any size [10]
[11], and possibly for adenomas in the 10- to 20-mm size range [12]
[13]
[14]. For lateral spreading adenomas ≥ 20 mm in size, cold resection is less effective
than resection using electrocautery [15]. The principal advantage of using cold resection is a very low risk of delayed hemorrhage
[16]. This may result from cold snare resection cutting through the submucosal plane
at a more superficial level compared with snare resection using electrocautery. The
same difference is the rationale for using electrocautery to remove lesions of any
size considered to have any significant risk of invasive cancer. However, the overwhelming
majority of colorectal lesions, particularly those ≤ 10 mm in size, can be predicted
to have essentially no risk of cancer based on their endoscopic features [17]. Because the risk of delayed hemorrhage is so low after resection of colorectal
lesions using cold techniques, there is no rationale for placement of prophylactic
clips after cold resection. However, we have anecdotally encountered individual physicians
with high rates of prophylactic clip closure after cold resection, including experts
who have used prophylactic clip closure after cold resection of large lesions, and
some physicians in our own group who use prophylactic clip closure after cold resection
of colorectal lesions regardless of size.
Unnecessary prophylactic clip closure increases the cost of colonoscopy because hemostatic
clips are expensive. Further, unnecessary clip closure adds time and inefficiency
to the colonoscopy procedure and creates unnecessary carbon emissions and plastic
waste. To assess the frequency of inappropriate clip closure in our own practice,
we performed an audit of clip use after cold resection during routine surveillance,
screening, and diagnostic colonoscopies in our academic endoscopy practice.
Methods
We performed a retrospective audit of clip usage after colorectal polyp resection
for 29 physicians performing colonoscopy in two university-based outpatient academic
endoscopy units associated with Indiana University. The audit interval began in January
2023 and extended into June 2023. The principal purpose of the audit was a quality
review of physician assignment of intervals for screening and post-polypectomy surveillance
colonoscopy, using the 2019 US Multi-society Task Force on Colorectal Cancer recommendations
[18]. A target of 100 consecutive patients undergoing routine screening, surveillance,
or diagnostic colonoscopy by each physician were reviewed. Both the number of cases
involving polyp resection and the number of polyp resections varied between endoscopists.
Patients were excluded if they had inflammatory bowel disease (IBD), inherited colorectal
cancer syndromes, serrated polyposis syndrome, or were referred to our center for
resection of a colorectal polyp.
All physicians were attending faculty members. Some physicians had fewer than 100
colonoscopies reviewed if their primary appointments were at the Veterans Administration
system hospital or the safety-net hospital in our system, or if their primary clinical
duties included less general endoscopy. The two outpatient centers included in the
study serve primarily private practice patient populations. Physicians were grouped
into the categories of advanced endoscopists, general gastroenterologists, IBD specialists,
liver specialists, or motility specialists, based on how they are identified within
our gastroenterology practice. The physician who performs the most endoscopic resections
of large colorectal lesions at our center (DKR) was classified as a general gastroenterologist.
Advanced endoscopists at our center perform primarily endoscopic ultrasound, endoscopic
retrograde cholangiopancreatography, per-oral endoscopic myotomy of the lower esophageal
sphincter and pylorus, and endoscopic submucosal dissection.
We used the colonoscopy reports identified in an electronic report generation system
database (Provation, Minneapolis, Minnesota, United States) to conduct the audit.
For each colonoscopy involving polyp resection, we recorded polyp size, location,
shape, method of resection, whether or not any clips had been placed post resection,
and if clips were placed, the number used. Lesions were grouped into three size categories:
≥ 20 mm, 10 to 19 mm, and 1 to 9 mm.
Statistical analysis
The number of colonoscopies targeted for each physician was arbitrary and selected
for the quality review of appropriate use of screening and post-polypectomy surveillance
intervals as a reasonable number to assess each physician’s adherence to the 2019
US Multi-society Task Force on Colorectal Cancer recommendations.
In general, statistics were descriptive. We used Jeffrey’s binomial procedure to calculate
confidence intervals for clip use proportion. All calculations were performed using
SPSS version 29. We used chi-square test or Fisher’s exact test, as appropriate, to
compare clip use for lesions removed by cold resection in patients on no anticoagulation
compared with those on aspirin only or with those on either anticoagulation or a non-aspirin
antiplatelet agent. Statistical significance was set at 0.05.
Results
There were 29 endoscopists who performed routine screening, surveillance, and diagnostic
colonoscopies during the audit interval, including four advanced endoscopists, 11
general gastroenterologists, three IBD specialists, five liver specialists (hepatologists),
and six motility specialists. All 29 endoscopists removed some lesions in the 1- to
9-mm category, 25 removed at least one lesion in the 10- to 19-mm category, and 17
removed at least one lesion ≥ 20 mm.
[Table 1] shows the number of lesions in each of the three size categories removed by cold
resection, and the number in which one or more hemostatic clips were placed. There
were 3784 total lesions evaluated in the audit, of which 3565 lesions were 1 to 9
mm in size ([Table 1]). The table also shows the range of lesions followed by clip placement for all endoscopists,
and with the breakdown by subspecialty.
Table 1 Clip use in cold resections by size.
|
CI, confidence interval; IBD, inflammatory bowel disease.
|
|
≥ 20 mm
|
Physician category
|
Number of physicians
|
Number of procedures
|
Number of lesions
|
Percentage of resections with clips used
(95% CI)
|
Range of clip use by individual physicians
|
|
All physicians
|
7
|
11
|
12
|
41.7% (18.0%-68.8%)
|
0.0%-100.0%
|
|
10–19 mm
|
Physician Category
|
Number of physicians
|
Number of procedures
|
Number of lesions
|
Percentage of resections with clips used
(95% CI)
|
Range of clip use by individual physicians
|
|
All physicians
|
23
|
164
|
207
|
19.3% (14.4%-25.1%)
|
0.0%-100.0%
|
|
Advanced endoscopists
|
3
|
13
|
17
|
35.3% (16.3%-58.9%)
|
28.6%-66.7%
|
|
General
|
9
|
88
|
109
|
16.5% (10.5%-24.3%)
|
0.0%-60.0%
|
|
IBD
|
3
|
29
|
41
|
7.3% (2.1%-18.3%)
|
0.0%-25.0%
|
|
Liver
|
3
|
14
|
18
|
27.8% (11.5%-50.6%)
|
0.0%-30.8%
|
|
Motility
|
5
|
20
|
22
|
36.4% (18.9%-57.1%)
|
0.0%-100.0%
|
|
1–9 mm
|
Physician category
|
Number of physicians
|
Number of procedures
|
Number of lesions
|
Percentage of resections with clips used
(95% CI)
|
Range of clip use by individual physicians
|
|
All physicians
|
29
|
1307
|
3565
|
2.8% (2.3%-3.4%)
|
0.0%-45.0%
|
|
Advanced endoscopists
|
4
|
75
|
194
|
1.5% (0.4%-4.1%)
|
0.0%-3.0%
|
|
General
|
11
|
628
|
1653
|
0.9% (0.5%-1.5%)
|
0.0%-6.3%
|
|
IBD
|
3
|
192
|
480
|
1.3% (0.5%-2.6%)
|
0.0%-3.9%
|
|
Liver
|
5
|
185
|
583
|
3.9% (2.6%-5.8%)
|
0.0%-25.5%
|
|
Motility
|
6
|
227
|
655
|
8.2% (6.3%-10.5%)
|
0.0%-45.0%
|
Three individual physicians, including two motility specialists and one liver specialist,
used clips in a large percentage of lesions after cold resection. For example, for
1- to 9-mm lesions, these three physicians placed clips in 18.8%, 25.5%, and 45.0%
of lesions after cold resection. These three physicians accounted for 8.1% of all
cold resections in the 1- to 9-mm size range but accounted for 62.4% of lesions 1
to 9 mm with clip placement. At the other end of the usage spectrum, there were 16
physicians who utilized clips in ≤ 1% of lesions 1 to 9 mm in size.
There were 879 lesions resected cold in patients on or resuming antiplatelet agents
or
anticoagulants shortly post-procedure. Of these 530 (60.3%) were using aspirin only,
192
(21.8%) were on anticoagulation alone (warfarin or direct-acting oral anticoagulant),
51
(5.8%) were on dual antiplatelet therapy, 46 (5.2%) were on non-aspirin antiplatelet
therapy
alone, 37 (4.2%) were on aspirin plus anticoagulation, 13 (1.5%) were on non-aspirin
antiplatelet therapy plus anticoagulation, and 10 (1.1%) were on two anticoagulant
medications. There were no significant differences in clip use after cold resection
of lesions
in patients on no blood thinners of any type compared with those on aspirin only,
either for
all physicians combined or any of the physician subspecialty groups ([Table 2]). Clip use after cold resection for all lesions was higher in patients on
anticoagulation or non-aspirin antiplatelet agents compared with those on no blood
thinners
for all physicians combined (5.7% vs 3.5%; P = 0.039), and among
the advanced endoscopists (12.8% vs. 1.5%; P = 0.005), but there
were no significant differences between these groups of lesions among the four other
physician
subspecialty groups ([Table 2]). The three physicians with the highest clip use had clip use rates after cold
resection of lesions 1 to 9 mm in size in patients on no anticoagulation or antiplatelet
agents (including aspirin) of 19.6%, 26.3%, and 46.2%, and each physician independently
had
higher clip use in these lesions compared with the other 26 physicians combined (P < 0.001 for each of the three physicians).
Table 2 Clip use after cold resection among all physicians and within physician groups when
no anticoagulation or antiplatelet agent was in use, when aspirin only was in use,
and when
anticoagulation or an antiplatelet agent was used or resumed shortly
post-procedure.
|
Physician category
|
No anticoagulation or antiplatelet agent
|
On aspirin only
|
On or resuming anticoagulation or non-aspirin antiplatelet agent
|
|
* Lesions clipped/all lesions in category (%).
IBD, inflammatory bowel disease.
|
|
All physicians
|
*102/2905 (3.5)
|
24/530 (4.5)
|
20/349 (5.7)
|
|
Advanced endoscopists
|
2/130 (1.5)
|
1/34 (2.9)
|
6/47 (12.8)
|
|
General
|
25/1420 (1.8)
|
3/187 (1.6)
|
6/159 (3.8)
|
|
IBD
|
7/442 (1.6)
|
3/53 (5.7)
|
1/32 (3.1)
|
|
Liver
|
20/434 (4.6)
|
3/108 (2.8)
|
5/59 (8.5)
|
|
Motility
|
48/479 (10.0)
|
14/148 (9.5)
|
2/52 (3.8)
|
During the audit interval, there were an additional 121 lesions removed during routine
screening, surveillance, or diagnostic examinations using electrocautery. Thus, 3.1%
(121 of 3905) of all colorectal lesions removed during these routine colonoscopies
utilized electrocautery. For all lesions removed, the rate of clip placement for lesions
removed after cold resection was 3.9% (146 of 3784), compared with 71.1% (86 of 121)
of the lesions removed by electrocautery. Of the 86 lesions for which clips were placed
after resection with electrocautery, seven fulfilled the traditional criteria for
clip closure of size ≥ 20 mm and location proximal to the splenic flexure. There were
57 lesions in which clip placement was used after electrocautery that were pedunculated
or semi-pedunculated in shape. Four lesions involved treatment of a recurrence using
electrocautery. In the remaining 18 lesions, there was no clinically evident basis
for prophylactic clip closure, except that for six of these 18 lesions, the patient
was scheduled for reinstitution of anticoagulation or antiplatelet agents. The average
number of clips placed per lesion after electrocautery was 1.72, which was similar
to the average 1.68 clips placed per lesion when clips were used after cold resection.
The average number of clips placed per lesion after cold resection of lesions measuring
1 to 9 mm was 1.53. Overall, there were 246 clips used on cold resection sites, and
148 clips used on sites after use of electrocautery, so that 62.4% of all clips used
in the audit were on cold resection sites.
Discussion
In this report, we describe an audit of clip use after resection of colorectal lesions
in patients undergoing routine screening, surveillance, and diagnostic colonoscopy
in two outpatient endoscopy centers. Importantly, patients who were referred for resection
of colorectal lesions were excluded from the study. Many of these patients referred
for resection had lesions that would fulfill the standard criteria for prophylactic
clip closure, including size ≥ 20 mm, location proximal to the splenic flexure, and
removal by electrocautery.
Thus, these results reflect clip use in patients undergoing routine colonoscopy, almost
entirely without preexisting knowledge of whether and what colorectal lesions were
present. Several important findings were generated from the audit, which led to direct
feedback to the entire group of endoscopists, and specific feedback to heavy users
of clip placement after cold resection. First, specific individuals were using prophylactic
clip closure more frequently than expected and more commonly than the other physicians
for lesions with a negligible risk of delayed post-polypectomy hemorrhage [16], and in patients not on any anticoagulation or antiplatelet agents. Available evidence
suggests that clip placement after cold snare resection is not helpful even when anticoagulation
is continued or resumed [19]. Each of the physicians overusing clips responded favorably to feedback and instruction
about appropriate use of clips for prophylaxis and agreed to change their practice.
Additional audit of their practice is expected. Second, while cold resection was overwhelmingly
the most common method of resection identified during the audit (> 96.9% of resections),
and while clip use was much less common after cold resection than after resection
with electrocautery, clip use after cold resection still accounted for the bulk of
clip use during routine screening, surveillance, and diagnostic colonoscopies. Regarding
pedunculated polyps, it is our institutional practice to place clips after resection
rather than using clips or ligature devices prior to resection in order to maximize
the oncologic resection margin in cases of pedunculated polyps harboring malignancy
[20]. Clips generate a variable amount of waste depending on manufacturer, but in general
lead to higher carbon emissions than either snares or forceps [21]. Therefore, more appropriate use of clips after cold resection could reduce endoscopy
unit costs, improve the efficiency of colonoscopy for specific over-users of clips,
and reduce carbon emissions and plastic waste associated with device use.
Third, the number of clips used per resection site in the audit was similar after
cold resection versus resection with electrocautery. This suggests that endoscopists
using clips after cold resection were often trying to close the entire cold resection
site. The senior author for this study recommends that clip placement after cold resection,
regardless of the lesion size, should only be utilized for instances of persistent
active immediate bleeding. Prior to clip placement, simple measures such as: 1) use
of the water jet to create a submucosal cushion and tamponade; 2) direct pressure
on the bleeding area with the scope tip; and 3) reopening the snare and re-grasping
the bleeding point and applying direct tamponade with the snare without re-transection
of the submucosal defect will usually stop immediate bleeding without the need for
clip placement. If clip placement is required, only the number of clips needed to
stop immediate bleeding (usually one clip) is needed, and there is no rationale for
closing the entire defect. Previous reports have also noted that immediate bleeding
can be prevented to some degree by squeezing the ensnared tissue for a few seconds
before transection [22]. Anecdotally, this seems to create less immediate bleeding than rapid cold transection.
Limitations of the study include its retrospective nature. However, retrospective
analysis can accurately identify important practice patterns, as was the case in this
study. The most significant limitation is the single academic center design. Excessive
clip use may or may not occur at other centers. Neither endoscopy unit in our study
is an ambulatory surgery center (ASC). In an ASC, we would expect financial pressures
to limit excessive hemostatic clip use. Such pressures were not present for endoscopists
in this study. Our study only suggests that audit of appropriate clip use may be helpful
in some settings.
Conclusions
In summary, we found through an audit of clip usage that in an academic medical center,
a
significant fraction of clip usage occurred after cold transections, and significant
evidence
that the majority of this clip usage was inappropriate and unnecessary. This audit
allowed
feedback that should lead to a reduction in cost, improved efficiency for certain
operators,
and a reduction in carbon emissions and plastic waste associated with device use.
Although
this problem might be exclusive to our center, this seems unlikely. We suggest that
others
evaluate and report patterns of clip use in their practice, and that professional
societies
emphasize current concepts about appropriate use of prophylactic clips.
