Introduction
Colorectal cancer (CRC) is an important cause of morbidity and mortality worldwide
[1 ]. CRC screening programs reduce CRC incidence [2 ], as the endoscopic removal of adenomas prevents their progression to CRC [3 ]
[4 ]
[5 ].
Hot snare polypectomy (HSP) has been the standard of care for the resection of lesions
over 5 mm. However, cold snare polypectomy (CSP) has recently grown in popularity,
as it presents a low risk of complications [6 ]
[7 ], even in patients receiving anticoagulant therapy [8 ]. Furthermore, CSP shortens the procedure time [9 ]
[10 ]. Thus, the recent European Society of Gastrointestinal Endoscopy clinical guidelines
recommend CSP for removal of polyps ≤ 5 mm and suggest CSP for flat/sessile lesions
6–9 mm [11 ].
There is contradictory evidence on the risk of residual polyp, and therefore an optimal
assessment of the resection margin is of paramount importance. A pilot study evaluating
CSP resection of 6–9 mm polyps showed important rates of retrieval failure and inadequate
histopathological evaluation of the horizontal margins [6 ]. Similarly, a prospective observational study concluded that the lateral margins
of 67.1 % of polyps < 10 mm resected by CSP were inadequate for assessment [12 ]. Thus, evaluation of biopsies from the margins of the mucosal defect is the preferred
method for determination of the complete resection rate of CSP [12 ]
[13 ].
A recent multicenter clinical trial conducted in Japan, which employed biopsies from
the mucosal defect margin to compare complete resection rates, demonstrated that CSP
is not inferior to HSP (98.2 % vs. 97.4 %) [14 ]. This result is supported by previous studies of lower statistical power [15 ]
[16 ].
Studies comparing the impact of CSP and HSP on post-colonoscopy symptoms are lacking
however. A single-center trial reported a higher proportion of overall abdominal symptoms
in patients undergoing HSP (20 % vs. 2.5 %) [16 ], but the abdominal pain assessment was dichotomous, without any further information
about its temporal evolution.
Therefore, our aim was to compare HSP and CSP for small polyps in terms of complete
resection and to assess the association between the polypectomy technique and adverse
events, namely post-procedural abdominal pain, and intraprocedural and delayed bleeding.
Methods
The study was a multicenter, prospective, randomized clinical trial comparing two
endoscopic polypectomy techniques in colorectal polyps of 5–9 mm in size. The study
was conducted between February and November 2019 at seven Spanish centers, including
primary, secondary, and tertiary hospitals. Participating endoscopists had a minimum
experience of 300 colonoscopies per year. The institutional review boards of all centers
approved the study protocol between October 2017 and February 2019, and all patients
provided written informed consent.
Patient selection
Patients over the age of 18 years referred for colonoscopy for any indication (positive
fecal occult blood test, post-polypectomy follow-up, CRC family history/other screening,
previous CRC, or symptoms) were prospectively invited to participate. Exclusion criteria
for randomization included pregnancy and polypectomy contraindication owing to continuation
of anticoagulant/antithrombotic agents (except aspirin) or uncorrected severe coagulopathy/thrombocytopenia.
Those who were found to have one or more polyps measuring 5–9 mm were enrolled. All
5–9 mm lesions were considered eligible, regardless of morphology or pit pattern classification.
After enrollment, polyps lost for histological analysis or those presenting normal
mucosa in the histological analysis of the polypectomy specimen were further excluded.
Randomization and concealment
Enrolled patients were assigned in a 1:1 ratio to the CSP or HSP group using a computer-generated
random sequence. Randomization was stratified by institution, so each center received
a unique set of completely opaque, sequentially numbered envelopes containing the
assignments. After identifying the first 5–9 mm polyp, an envelope was opened. Once
the polypectomy technique was assigned, it was employed in all 5–9 mm lesions in that
patient.
Intervention
All colonoscopies were performed by trained endoscopists with appropriate preparation
and sedation according to the usual clinical practice at each center. Standard video
colonoscopes were used. High definition, magnification, and image enhancement functions
were not mandatory.
Polyp size was estimated using an open snare as reference. The morphology was defined
by the Paris classification [17 ]. Once eligibility was confirmed, the polyp was removed by HSP or CSP as randomly
determined. The type of snare was freely chosen by the endoscopist in the HSP group
from the following devices: 25 mm round snare (Olympus, Barcelona, Spain); 13, 15,
20, and 33 mm round snares (Boston Scientific, Marlborough, Massachusetts, USA); 15 mm
asymmetric snare (MTW, Wesel, Germany); 15 mm round snare (Cook Medical, Bloomington,
Indiana, USA). In the CSP group, in addition to all previously mentioned snares, CSP-dedicated
10 mm snares (Boston Scientific) could also be employed. Submucosal injection was
not permitted.
In the CSP group, the technique used was cold resection of the polyp without tenting,
followed by suction of the transected polyp. In the HSP group, an electrocoagulation
unit (Erbe Elektromedizin, Tübingen, Germany) was used in the Endocut mode. When a
polyp could not be removed using the CSP technique, HSP rescue was permitted. Polyps
under 5 mm or over 9 mm were treated according to usual clinical practice.
After resection, the mucosa was carefully observed. When residual polyp tissue was
clearly recognized, additional removal using the same snare was allowed. After confirming
the absence of residual polyp tissue by endoscopic inspection, random biopsies from
the right and left edges of the margins of the mucosal defect were performed ([Fig. 1 ]). Directed biopsies were also permitted in case-specific areas of the marginal mucosa
that had a suspicious appearance (e. g. wrinkled folds, distorted pit pattern).
Fig. 1 Study procedure for cold snare polypectomy (CSP). a A sessile polyp detected in the colon. b The Boston Scientific 10-mm polypectomy snare was used for CSP. The size of this
polyp was estimated to be 6 mm. c CSP was performed. d Biopsy samples were taken from two marginal sites located symmetrically on the left
and right of the mucosal defects to determine the presence or absence of residual
polyp tissue.
Endoscopic hemostasis was carried out when active hemorrhage continued for ≥ 60 seconds.
Preventive hemostasis, defined as prophylactic coagulation of vessels or red spots
in the ulcer or clipping of a nonbleeding post-polypectomy mucosal defect, was not
allowed.
Patients received and were trained to complete a questionnaire after the procedure.
The questionnaire included items on the evolution of abdominal pain and other possible
adverse events.
Post-procedure monitoring
All patients were contacted via a centralized telephone call 21–28 days after the
colonoscopy by a single experienced research nurse who was blinded to procedure allocation.
Patients underwent a standardized interview regarding possible adverse events to retrieve
the abdominal symptomatology data from the questionnaire. For patients in whom adverse
events had required medical evaluation, local investigators were contacted and asked
to submit a report. The endoscopic finding of active bleeding, adherent clots or visible
vessels confirmed the source of bleeding in post-colonoscopy hemorrhages. In patients
with multiple polypectomies where none of the previously stated findings could be
found, the largest 5–9 mm polyp was considered the probable source of bleeding, regardless
of the number and size of polypectomies.
Histological analysis
All polypectomy specimens were evaluated at each center according to its usual protocol.
Biopsies of the mucosal defect were centrally assessed by a single pathologist who
had more than 5 years’ experience in digestive pathology (J.F.) and was blinded to
the histological diagnosis of the lesion and the polypectomy technique.
Outcomes and definitions
The primary end point of the study was the complete resection rate. Complete resection
was defined as the presence of normal mucosa or burn artifacts in the biopsies from
the margins of the mucosal defect.
Secondary end points included the assessment of adverse events (intraprocedural and
delayed bleeding, post-procedure abdominal pain, and other minor adverse events) and
predictive factors associated with incomplete resection. Randomization of patients
instead of polyps was required to assess delayed bleeding and post-procedure abdominal
pain. Intraprocedural bleeding was defined as spurting or oozing that continued after
60 seconds of observation without continuous washing. Bleeding that ceased within
the 60-second observation time was not labelled as an adverse event but was instead
recorded as a self-limited bleed. Delayed hemorrhage was defined as rectal bleeding
between discharge and the telephone follow-up contact. Severity of adverse events
was defined according to the American Society for Gastrointestinal Endoscopy recommendations
[18 ]. The evolution of abdominal pain was self-assessed at 1, 3, and 5 hours after the
endoscopic procedure using a visual analog scale, which was later categorized as follows:
0 absent, 1–3 mild, 4–6 moderate, 7–10 severe. The first day with absence of pain
was also recorded.
The type of polypectomy was defined according to how the polypectomy was performed,
regardless of the assigned group. The following categories were included: CSP en bloc
resection in a single maneuver; CSP en bloc resection with snare relocation (maneuvers
to relocate the snare were needed to complete the polypectomy); piecemeal CSP; HSP
en bloc resection; and piecemeal HSP.
Failed CSP was defined as those polypectomies randomized to CSP that could not be
completed and that were finally performed with a hot snare.
Sample size calculation
Sample size estimation was carried out using Stata software (StataCorp. 2013, College
Station, Texas, USA). We anticipated a CSP incomplete resection rate of 10 %, according
to the 6.5 %–22.7 % rates reported by Pohl et al. for individual endoscopists [19 ] and the 8.5 % reported in a previous study [20 ]. To detect a 7 % difference, with an α risk of 5 % and 80 % statistical power, a
total of 315 lesions per group were deemed necessary. Assuming a conservative 20 %
proportion of losses (polyps not recovered for histological assessment, failed CSPs,
normal mucosa on histological analysis), we estimated 394 lesions/group would be required.
Data retrieval and statistical analysis
Data were collected and managed using the Spanish Digestive Endoscopy Society Research
Electronic Data Capture tool. This is a secure, web-based application created to support
data capture for research studies providing semi-automatic data quality control [21 ].
The statistical analysis was carried out using Stata software (StataCorp. 2013). Continuous
variables were summarized using mean and standard deviation or median and interquartile
range (IQR) for data with non-normal distribution according to the Kolmogorov–Smirnov
test; categorical variables were expressed as percentages. Baselines characteristics
were compared using the chi-squared test and t test, as appropriate.
The difference in the proportions of incomplete polypectomies (primary outcome) and
adverse events was assessed using logistic regression models based on the generalized
estimating equation considering intrasubject correlation. The proportion of patients
with abdominal pain in both groups was compared using the Z test of homogeneity without the Yates correction. The analysis was carried out by
intention to treat, regardless of the type of polypectomy finally performed.
To evaluate the factors associated with incomplete polypectomies and those associated
with the presence of abdominal pain, multivariable logistic regression techniques
were used. Incomplete polypectomy analysis was adjusted for cluster effect; variables
first assessed by univariable analysis included: age, sex, working shift, time of
the procedure (initial half vs. final half of the working shift), indication, procedure
duration, total number of polyps and of 5–9 mm polyps, order of resection among 5–9 mm
polyps, type of snare (size), polypectomy technique (CSP vs. HSP), type of polypectomy
(en bloc vs. piecemeal), polyp size, location, and morphology, endoscopist experience
in years, and endoscopist (including only endoscopists with ≥ 20 lesions, divided
into terciles according to their individual complete resection rates). Risk factors
for abdominal pain were assessed only in patients whose polyps were all 5–9 mm and
excluded patients who also had polyps < 5 mm or > 9 mm. Factors included age, sex,
indication, procedure duration, total number of 5–9 mm polyps, polypectomy technique,
endoscopist, and gas employed for insufflation. Predictors with P < 0.10 in univariable analysis were then evaluated in multivariable logistic regression
models to determine the adjusted odds ratios (OR).
Results
A total of 1538 patients were recruited between February and November 2019 across
7 centers by 20 endoscopists. Recruitment ended when the predetermined sample size
was reached. As shown in [Fig. 2 ], 1042 patients were excluded, mostly because they did not have any 5–9 mm polyps.
From 496 patients randomized (237 to CSP and 259 to HSP), 8 patients (5 in CSP group
and 3 in HSP) were further excluded owing to retrieval failure of at least one of
their polypectomy specimens. Thus, the CSP group included 232 patients with 394 polyps,
and the HSP group included 256 patients with 397 lesions.
Fig. 2 Study flow chart.
Baseline data
An overview of patient background characteristics and procedures is summarized in
[Table 1 ]. Carbon dioxide insufflation was used in 96.6 % of patients in the CSP group and
in 94.1 % of patients in the HSP group. Both groups presented a median procedure time
of 25 minutes (P = 0.85); this lack of differences was also observed in a further analysis limited
to patients presenting only 5–9 mm lesions (21.2 minutes with CSP vs. 22.1 minutes
with HSP; P = 0.34). Among the 488 patients included, 211 (43.2 %) presented one 5–9 mm polyp,
91 (18.6 %) presented 2 polyps, 37 (7.6 %) presented 3, and the remaining 149 patients
(30.5 %) presented more than 4 lesions.
Table 1
Patient characteristics.
CSP (n = 232)
HSP (n = 256)
P
Age, median (IQR), years
64.7 (56.7–70.5)
64.5 (57–70.7)
0.78
Male sex, n (%)
150 (64.7)
169 (66.0)
0.75
Pharmacological treatment, n (%)
28 (12.1)
17 (6.6)
0.04
3 (1.3)
6 (2.3)
0.51
7 (3.0)
8 (3.1)
0.95
8 (3.5)
6 (2.3)
0.47
Indication, n (%)
0.26
86 (37.1)
74 (28.9)
43 (18.5)
71 (27.7)
28 (12.1)
34 (13.3)
12 (5.2)
8 (3.1)
63 (27.2)
69 (27.0)
Colonoscopy duration, median (IQR), minutes
25 (20–32)
25 (20–33)
0.85
Insufflation method, n (%)
0.21
224 (96.6)
241 (94.1)
8 (3.5)
15 (5.9)
Total number of polyps, n (%)
0.71
156 (67.2)
183 (71.5)
41 (17.7)
43 (16.8)
25 (10.8)
19 (7.4)
10 (4.3)
11 (4.3)
CSP, cold snare polypectomy; HSP, hot snare polypectomy; ASA, acetylsalicylic acid;
NOAC, novel oral anticoagulant; FOBT, fecal occult blood test; CRC, colorectal cancer;
IQR, interquartile range.
Overall, 394 lesions were included in the CSP group and 397 were included in the HSP
group. Polyp characteristics are summarized in [Table 2 ]. Background characteristics of the allocated polyps (location, morphology, histology,
and size) were comparable between the two groups. One lesion (0.1 %) with superficial
(< 1000 µm) submucosal cancer invasion was found in a 9 mm pedunculated polyp resected
with cold snare; biopsies confirmed complete resection.
Table 2
Polyp characteristics.
CSP (n = 394)
HSP (n = 397)
P
Size, median (IQR), mm
6 (5–7)
6 (5–7)
0.92
Morphology, n (%)
0.50
252 (64.0)
235 (59.2)
14 (3.6)
19 (4.8)
24 (6.1)
32 (8.1)
93 (23.6)
102 (25.7)
11 (2.8)
8 (2.0)
0
1 (0.3)
Location, n (%)
0.79
111 (28.2)
106 (26.7)
23 (5.8)
22 (5.5)
65 (16.5)
64 (16.1)
10 (2.5)
12 (3.0)
46 (11.7)
49 (12.3)
92 (23.4)
108 (27.2)
47 (11.9)
36 (9.1)
Histology, n (%)
0.29
244 (61.9)
250 (63.0)
25 (6.4)
38 (9.5)
1 (0.3)
0
65 (16.5)
52 (13.1)
16 (4.1)
9 (2.3)
34 (8.6)
34 (8.6)
9 (2.3)
14 (3.5)
Directed biopsies, n (%)
68 (17.3)
43 (10.8)
0.01
Type of polypectomy, n (%)
n/a
328 (83.2)
0
45 (11.4)
0
17 (4.3)
0
3 (0.8)
384 (96.7)
1 (0.3)
13 (3.3)
CSP, cold snare polypectomy; HSP, hot snare polypectomy; IQR, interquartile range;
SSA/P, sessile serrated adenoma/polyp.
Polypectomy description
Among the 394 lesions found in the CSP group, rescue HSP was required in 4 (1.0 %)
failed CSPs. All 397 polyps assigned to HSP were resected with this technique, although
13 (3.3 %) underwent a piecemeal HSP. Dedicated snares were employed in 187 lesions
(47.5 %) of the CSP group. Directed biopsies were performed more frequently in the
CSP group than in the HSP group (17.3 % vs. 10.8 %; P = 0.01).
Resection rates
A total of 394 polyps from 232 patients in the CPS group and 397 polyps from 256 patients
in the HSP group were retrieved. Overall, 16 polypectomy samples showing normal colonic
mucosa (5 CSP and 11 HSP) and 3 lesions whose marginal biopsy containers were lost
(2 CSP and 1 HSP) were excluded from the resection rate analysis; thus, 387 polyps
in the CSP group and 385 in the HSP group were finally analyzed. We observed no differences
in resection rates between the two groups: CSP 358 /387 (92.5 %) vs. HSP 362/385 (94.0 %);
difference 1.5 %, 95 % confidence interval [CI] –1.9 % to 4.9 % (P = 0.38).
Polyps undergoing directed biopsy had an incomplete resection rate of 13.1 % (8/61)
with CSP and 14.0 % (7/50) with HSP (P = 0.88), whereas in polypectomies undergoing random biopsies, incomplete resection
rates were 6.1 % (21/342) and 4.7 % (16/340), respectively (P = 0.46).
Among the CSP procedures, dedicated 10 mm cold snares did not improve complete resection
rates compared with conventional snares: 95.7 % (67/70) vs. 91.4 % (661/723), respectively
(P = 0.20). Among endoscopists who performed ≥ 20 polypectomies, individual incomplete
resection rates ranged from 1.5 % (1/65) to 13.9 % (5/36). Univariable and multivariable
analysis assessing risk factors for incomplete resection are shown in [Table 3 ].
Table 3
Univariable and multivariable odds ratios for risk factors of incomplete resection[1 ].
Univariable analysis
Multivariable analysis
OR (95 %CI)
P value
aOR (95 %CI)
P value
Female sex
1.78 (1.04–3.05)
0.04
2 (1.10–3.64)
0.02
Endoscopist[2 ]
2.21 (1.60–3.05)
< 0.001
2.57 (1.83–3.59)
< 0.001
Piecemeal polypectomy
3.78 (1.50–9.55)
0.01
4.27 (1.47–12.40)
0.01
Nonadenomatous polyps
1.99 (1.14–3.46)
0.02
2.02 (1.07–3.82)
0.03
Resection order number (first as reference)
1.21 (1.00–1.48)
0.05
1.38 (1.09–1.74)
0.01
Flat polyps
0.39 (0.18–0.87)
0.02
Sessile polyps
1.94 (1.03–3.68)
0.04
8–9 mm polyp size
1.84 (1.01–3.35)
0.05
(a)OR, (adjusted) odds ratio; CI, confidence interval.
1 Only variables reaching a P value < 0.10 in the univariable analysis are presented in the table. Other variables
assessed by univariable analysis were: patient age, working shift, time of the procedure
(initial half vs. final half of the working shift), indication, procedure duration,
total number of polyps and of 5–9 mm polyps, type of snare, polypectomy technique
(cold vs. hot snare), polyp location, and endoscopist experience in years).
2 Individual endoscopist performance according to terciles of complete endoscopic resection;
best tercile as reference.
Adverse events
Overall adverse events were identified in 58 (25.0 %) of the 232 CSP patients and
in 75 (29.3 %) of the 256 HSP patients (P = 0.29).
Intraprocedural bleeding was observed in only 3/394 CSP procedures (0.8 %; 2 Is and
1 Ip polyp) and 7/397 HSP procedures (1.8 %; 2 Ip, 1 Isp, and 4 Is lesions) (P = 0.34); successful endoscopic clipping was applied in all cases. Self-limited bleeding
was observed in 58 CSP procedures (14.7 %) and in 17 HSPs (4.3 %; difference 10.4 %,
95 %CI 6.4 % to 14.5 %; P < 0.001).
Only two patients presented delayed hemorrhage at 24 and 48 hours after the procedure,
respectively, and both of them required hospitalization (one 5 mm sessile tubular
adenoma in the CSP group and one 9 mm semi-pedunculated tubular adenoma in the HSP
group). In another 36 patients (17 in the CSP group and 19 in the HSP group; P = 0.97), minimal post-colonoscopy rectal bleeding was recorded, which did not warrant
medical attention.
The abdominal pain analysis was performed only in patients whose polyps were all 5–9 mm
and excluded patients who also had polyps < 5 mm or > 9 mm; thus, 204 patients (101
in the CSP group and 103 in the HSP group) were included. Ambient air insufflation
was employed in two patients (2.0 %) in the CSP group and in four patients (3.9 %)
in the HSP group (P = 0.68). The evolution of abdominal pain is shown in [Fig. 3 ]. At 1 hour after the procedure, both groups presented similar proportions of symptomatic
patients (19/101 [18.8 %] with CSP vs. 19/103 [18.4 %] with HSP); however, at 5 hours
after the procedure, only 6/101 (5.9 %) presented symptoms in the CSP group (4.9 %
mild and 1 % moderate), whereas 17/103 patients (16.5 %) recorded pain in the HSP
group (12.6 % mild, 3.9 % moderate) (P = 0.02). In the CSP arm, 97.0 % (98 /101) of patients were asymptomatic 24 hours
after the procedure, whereas in the HSP group 86.4 % (89/103) were asymptomatic (P = 0.01), resulting in a 10.6 % (95 %CI 3.2 % to 18 %) decrease in the proportion
of symptomatic patients following CSP. Multivariable analysis identified HSP as the
only risk factor for post-procedure abdominal pain at 5 and 24 hours after the procedure
(OR 3.13, 95 %CI 1.18 to 8.30, P = 0.02 and OR 4.97, 95 %CI 1.38 to 17.7, P = 0.01, respectively).
Fig. 3 Temporal evolution of post-polypectomy abdominal pain. Pain was measured using a
visual analog scale and categorized into: absent (0), mild (1–3), moderate (4–6),
severe (7–10). Only patients without lesions < 5 mm or > 9 mm were included. CSP,
cold snare polypectomy; HSP, hot snare polypectomy.
Other minor adverse events were flatulence, reported in seven patients (3.0 %) in
the CSP group and in nine patients (3.5 %) in the HSP group (P = 0.75), and anal pruritus in four patients (1.7 %) and seven patients (2.7 %), respectively
(P = 0.45).
Discussion
Our multicenter randomized trial comparing CSP and HSP for 5–9 mm colorectal polyps
did not find significant differences regarding incomplete resection rates or post-polypectomy
bleeding. However, CSP was associated with a significant reduction in the intensity
and duration of post-procedural abdominal pain.
Incomplete resection rates found in our study resemble previously reported data. Pohl
et al. reported a 6.8 % incomplete resection rate for 5–9 mm nonpedunculated polyps,
with larger size and a diagnosis of sessile serrated adenoma/polyps (SSA/Ps) being
the strongest predictors of incomplete resection [19 ]. Other studies have reported incomplete resection rates ranging between 1.8 %–8.5 %
for CSP and 2.6 %–3.7 % for HSP [14 ]
[15 ]
[20 ]. The CRESCENT study reported an excellent 1.8 % for CSP and 2.6 % for HSP [14 ], but excluded SSA/Ps and hyperplastic polyps. Additional removal using the allocated
technique was also allowed, as in our design, but cold biopsy forceps could also be
employed. In our study, directed biopsies were more frequently performed in the CSP
group, maybe due to better differentiation of suspicious tissue in the absence of
burning artifacts. Other studies using post-polypectomy endoscopic mucosal resection
to assess incomplete resection rates report values of 3.4 % for CSP in 1–7 mm polyps
[13 ] and 3.9 % in 1–9 mm polyps [22 ].
Self-limited bleeding was more frequently encountered in the CSP group, but there
was no difference between the groups in rates of intraprocedural and delayed bleeding,
a finding that concurs with recent meta-analyses demonstrating the efficacy and safety
of CSP [23 ]
[24 ]. One patient in each group required hospitalization for delayed bleeding. Most previous
reports present delayed bleeding rates under 1 %, supporting these findings [10 ]
[14 ]
[15 ]
[16 ]
[20 ]
[25 ]. Interestingly, we observed these results despite not allowing prophylactic clipping
and with only three lesions (0.8 %) in the CSP group and seven (1.8 %) in the HSP
group receiving therapeutic clipping, whereas other studies report a proportion of
prophylactic/therapeutic clipping ranging between 1.8 %–35 % and 1.1 %–37 % for CSP
and HSP, respectively [14 ]
[20 ]
[25 ]. Our reduced use of clipping might be related to the 60-second observation period
set in our protocol, distinguishing inconsequential self-limited bleeding from intraprocedural
bleeds. This low rate of delayed bleeding in CSP supports previous findings demonstrating
the safety of this technique even in patients under continuous anticoagulant treatment
[26 ]. Moreover, our study included 56 pedunculated and 33 semi-pedunculated lesions,
which are usually excluded from CSP due to a theoretical higher risk of delayed bleeding.
The lack of differences observed supports the specific evaluation of CSP in 5–9 mm
Ip and Isp lesions in further studies.
Our study showed that CSP shortened and reduced the intensity of post-polypectomy
abdominal pain compared with conventional polypectomy, although it did not lead to
decreased health care utilization, extending the scarce available evidence supporting
this hypothesis [16 ]. It has been suggested that abdominal symptoms may be related to the number of polyps
removed or procedure duration [9 ]; in our study however, HSP was the only risk factor for pain at 5 and 24 hours after
the colonoscopy in patients whose polyps were all 5–9 mm. Abdominal pain and tenderness
within 12 hours after the polypectomy form the typical picture of the post-polypectomy
coagulation syndrome [27 ], presenting in 0.14 %–2 % of patients undergoing HSP [28 ]
[29 ]. We hypothesize that post-polypectomy coagulation syndrome represents a full spectrum
of disease, of which only the most severe cases are usually diagnosed. Thus, we consider
the increased number of symptomatic patients in the HSP group represent mild cases,
which in clinical practice do not seek medical care.
Although not a primary study aim, we found a wide range of incomplete resection rates
among endoscopists. This finding is consistent with the CARE study [19 ], where incomplete resection rates ranged from 6.5 % to 22.7 %. Female sex was also
associated with incomplete polyp resection. We hypothesize that in female anatomy,
which requires longer cecal intubation times [30 ], straightening the scope to an optimal position to visualize the lesion and adequately
ensnare a rim of normal mucosa might be more technically demanding.
Our study presents a series of strengths. We kept as close as possible to real clinical
practice. We included primary, secondary, and tertiary centers, snares were chosen
according to endoscopist preference and center availability, and all types of polyps,
regardless of optical diagnosis, morphology, and location were included. We consider
this heterogeneity allows us to better approach incomplete resection rates in a real-life
scenario. We also randomized patients instead of lesions, allowing the assessment
of post-procedure symptoms. Nevertheless, this study has certain limitations that
should be acknowledged. First, we evaluated residual tissue immediately after polypectomy.
Biopsy results obtained immediately after HSP may not be useful to predict recurrence
rates as samples include nonviable burnt tissue. Second, there are various possible
sources of bias: the endoscopist who performed the procedure also took the biopsy
samples and could choose the biopsy site; although the pathologist was blinded to
the procedure type, the burning effect of HSP is easily detectable. Despite these
limitations, our study provides important information regarding the usefulness and
safety of the CSP technique.
In conclusion, CSP is an effective and safe method that shortens and improves post-polypectomy
abdominal symptoms and should be recommended as the standard technique for 5–9 mm
colorectal polyps.
Efficacy and safety of cold versus hot snare polypectomy for small (5–9mm) colorectal
polyps: a multicenter randomized controlled trial
Acknowledgments
We gratefully thank our research nurse María Martínez Díez for her contribution.
POLIPEC HOT-COLD Study Group members: Javier Santos (Department of Gastroenterology,
Hospital Río Carrión, Palencia, Spain); Jesús Legido, Raúl Torres, Rebeca Amo (Department
of Gastroenterology, Hospital General de Segovia, Segovia, Spain); Javier Tejedor,
Sergio Bazaga (Department of Gastroenterology, Hospital Universitario Rio Hortega,
Valladolid, Spain); Aleida Miguel (Department of Gastroenterology, Hospital Universitario
Rio Hortega, Valladolid, Spain; Department of Gastroenterology, Hospital de Medina
del Campo, Valladolid, Spain).
