Impact of acetic acid chromoendoscopy on detection of serrated lesions in the proximal colon compared with white-light endoscopy.

Introduction

Colorectal cancer (CRC) is one of the most common malignancies worldwide and remains an important cause of morbidity and mortality
1
. Previously, colorectal tumors were believed to develop almost exclusively through the conventional adenoma-carcinoma sequence. However, after the serrated pathway was described, these precursor lesions are now recognized as responsible for up to 30% of CRC cases. Serrated lesions (SLs) comprise a heterogeneous group of lesions characterized by diverse morphological features, which are often subtle and easily overlooked during colonoscopy (
Fig. 1
). Several factors seem to contribute to the wide variability in SL detection rates observed among endoscopists, including bowel preparation quality, technical aspects of the procedure, morphological features, and personal expertise
2
3
. SLs are more often located in the proximal colon and may present as sessile (0-Is), slightly elevated (0-IIa), or flat (0-IIb) lesions, according to the Paris classification
4
. Their color is similar to the adjacent mucosa, with indistinct and irregular borders often covered by a mucus layer, which frequently contributes to missed detection
5
.

Robust epidemiological evidence suggests that the challenging diagnosis of SLs may partly explain their association with synchronous and metachronous colorectal neoplasia, as well as increased risk of advanced neoplasia linked to proximally located lesions
6
7
8
9
10
11
. With growing awareness of serrated polyps, there is increasing interest in identifying potential risk factors. Reported risk factors include smoking, alcohol consumption, obesity, high-fat diet, age, sex, type II diabetes, and race
12
13
14
15
16
17
.

Although high-definition white-light endoscopy (HD-WLE) remains the primary modality for diagnosing and evaluating colorectal lesions, Popoutchi et al.
18
demonstrated the usefulness of acetic acid chromoendoscopy (AAC) in the diagnosis and delineation of SLs, particularly in the right colon. Mucolytic properties of acetic acid enhance lesion visualization by removing adherent mucus and improving surface contrast (
Fig. 2
). In addition, mild changes in pH and ionic strength induced by acetic acid interfere with protein tertiary structure, altering their optical properties and resulting in transient epithelial whitening (acetowhite reaction)
19
20
. This effect increases surface opacity and masks the subepithelial vascular network, facilitating lesion recognition.

The acetowhite reaction enhances visualization of the pit pattern of colonic lesions, thereby facilitating interpretation, and disappears after acid neutralization, indicating a temporary effect
19
. Although previous studies have combined AAC with other image-enhanced techniques for lesion characterization, evidence specifically addressing the impact of AAC on overall SL detection rates remains limited.

The aims of this prospective study were to compare detection rates for SLs in the proximal colon between HD-WLE alone and HD-WLE combined with 2% AAC. In addition, the study aimed to assess the association between proximal SL detection and clinical risk factors.

Patients and methods

Study design and patient selection

This was a prospective, comparative study that analyzed patients who underwent colonoscopy at a tertiary center (Hospital Sírio-Libanês, São Paulo, Brazil) between January and July 2021. Based on expert consensus, an expected detection rate of approximately 5% for proximal serrated polyps using HD-WLE was assumed for sample size calculation. Accordingly, a sample of 434 patients provided 80% power with a two-sided alpha of 0.05. Eligible participants were adults over 18 years old undergoing colonoscopy, regardless of the indication for the procedure. Exclusion criteria included refusal to participate, history of inflammatory disease, hereditary polyposis or non-polyposis syndromes, previous right colon surgery, prior diagnosis of CRC, ongoing anticoagulation therapy or antiplatelet agents, as well as allergy to acetic acid or inadequate bowel preparation
21
. All participants provided written informed consent prior to enrollment, in accordance with institutional and national ethical standards and with the Declaration of Helsinki.

Procedure description
All patients underwent bowel preparation according to the institutional protocol, which consisted of a 10% mannitol solution combined with simethicone. Procedures were performed by experienced colonoscopists, each with a minimum of 10,000 diagnostic or therapeutic colonoscopies completed and an adenoma detection rate above 30% over the preceding year. A total of four colonoscopists participated in the study.
Colonoscopy was carried out under conscious sedation, administered by the anesthesiology team in all cases. All examinations were performed using high-definition Olympus CF-HQ190L colonoscopes coupled with the EVIS EXERA III CV-190 image processing system.

Examination of the proximal colon began with cecal intubation. The colonoscope was then withdrawn to the splenic flexure, which was defined as the anatomical landmark separating the proximal from the distal colon
22
23
24
25
26
. Under HD-WLE, any lesions or abnormalities identified during this first evaluation were treated or biopsied according to their clinical indication.

Subsequently, the colonoscope was reinserted into the cecum and an approximate total volume of 40 to 50 mL of a 2% acetic acid solution was instilled using an irrigation pump, enabling chromoendoscopy from the cecum to the splenic flexure. The proximal colon was then reexamined. All lesions identified during this second step were endoscopically resected and submitted to histopathological evaluation.
Use of antispasmodic medication was permitted on an as-needed basis at the discretion of the colonoscopist when peristaltic activity interfered with adequate inspection during withdrawal. Instillation of acetic acid did not result in clinically significant peristalsis that compromised mucosal evaluation.

Statistical analysis and data collection

Qualitative characteristics were summarized using absolute and relative frequencies, whereas quantitative characteristics were described using summary measurements (mean, standard deviation [SD], median, minimum, and maximum). The detection rate for SLs was reported for each method and compared using McNemar’s test. SL detection rates were further evaluated according to qualitative characteristics using chi-square tests, whereas quantitative variables were compared between groups with Student’s
t
-test or the Mann-Whitney test, as appropriate.

To evaluate the association between proximal serrated lesion detection and clinical characteristics, a multivariable logistic regression model was performed, with presence of a proximal serrated lesion as the dependent variable. Independent variables included age, sex, body mass index (BMI), systemic arterial hypertension, type 2 diabetes mellitus, dyslipidemia, smoking status, and bowel preparation quality. Variables were selected based on clinical relevance and prior univariate analyses. Associations were expressed as odds ratios with corresponding 95% confidence intervals (CIs), adopting a significance level of 5%.

Results

Sample characterization
Of 492 patients, 434 were included in the study. Among them, 159 were undergoing their first colonoscopy and 275 had previously undergone the procedure (36.6% and 63.4%, respectively). A total of 58 patients (11.7%) were excluded for the following reasons: 21 declined participation (36.2%), 10 had a diagnosis of inflammatory bowel disease (17.2%), three had a polyposis syndrome (5.1%), seven had a prior right colectomy (12%), five had a history and/or current treatment of CRC (8.7%), and 12 were using anticoagulant agents (20.7%).
Patient age ranged from 18 to 83 years (mean 50.1 years, SD 12.5). The majority were female (269, 62%), and 86 patients (19.8%) were classified as obese according to body mass index. Most participants (364, 83.9%) self-identified as White. Regarding indication for colonoscopy, 192 (44.2%) underwent the procedure for post-polypectomy surveillance, 146 (33.6%) for CRC screening, 36 (8.3%) for abdominal pain, 26 (6.0%) for unexplained diarrhea, and 34 (7.8%) for other reasons.

With respect to comorbidities, 87 patients (20.0%) were hypertensive, 45 (10.4%) diabetic, 65 (15.0%) dyslipidemic, and 46 (10.6%) were current smokers. Bowel preparation was classified using the Boston Bowel Preparation Scale as excellent/good in 414 patients (95.4%) and fair in 20 patients (4.6%) (
Table 1
).

Overall proximal lesion detection rates

A total of 340 proximal colonic lesions were identified in the study population. Of these, 207 lesions (60.9%) were detected during HD-WLE, whereas 133 lesions (39.1%) were detected only after AAC, representing a statistically significant difference favoring chromoendoscopy (
P
< 0.001) (
Table 2
).

The majority of proximal lesions detected during HD-WLE were located in the ascending colon, followed by the transverse colon and cecum.

Serrated lesion detection rates

Regarding proximal serrated lesions, a total of 71 SLs were identified in 71 patients. HD-WLE alone detected 26 SLs (36.6%), whereas 45 additional SLs (63.4%) were detected only after AAC, demonstrating a statistically significant increase in detection (P = 0.016) (
Table 2
).

Of the 133 additional proximal lesions identified after AAC, 45 (33.8%) were classified as SLs. These additional SLs were all < 10 mm, with most (60.0%) measuring between 1 and 5 mm. According to the Paris classification, the majority were flat or slightly elevated lesions (types 0-IIa or 0-Is) and were predominantly located in the transverse colon (53.3%), followed by the ascending colon (35.6%) and cecum (11.1%). Histopathological analysis showed that sessile SLs (SSLs) accounted for 68.9% of cases, hyperplastic lesions for 28.9%, and SSLs with dysplasia for 2.2%. No traditional SLs were detected among the additional lesions identified after AAC (
Table 3
).

Among the SLs detected during HD-WLE, 12 (46.2%) were located in the transverse colon, 10 (38.5%) in the ascending colon, and 4 (15.4%) in the cecum. Of the SLs detected after chromoendoscopy, more than half were located in the transverse colon, whereas the ascending colon and cecum accounted for 35.6% and 11.1%, respectively.

All 45 SLs detected after AAC measured < 10 mm, with 27 (60.0%) ranging from 1 to 5 mm. SSLs accounted for 31 cases (68.9%), hyperplastic lesions for 13 (28.9%), and SSLs with dysplasia for one case (2.2%). No traditional SLs were identified. A detailed description of SL characteristics detected after chromoendoscopy is provided in
Table 3
.

SLs and associated risk factors

In unadjusted analyses, none of the evaluated clinical variables were significantly associated with proximal SL detection (
P
> 0.05). However, in the multivariable logistic regression model, systemic arterial hypertension remained independently associated with proximal SL detection. Hypertensive patients had a 2.28-fold higher likelihood of harboring proximal SLs compared with non-hypertensive patients (95% CI 1.08–4.83; P = 0.032) (
Table 4
).

Discussion
Our study demonstrates that HD-WLE combined with 2% AAC can increase the detection rate for SLs in the proximal colon. When applied sequentially during the same examination, AAC was associated with a significant increase in SL detection compared with HD-WLE alone. To date, this is one of the few prospective studies reporting improved proximal SL detection using this technique in a real-world clinical setting.

SLs represent a major technical challenge due to their subtle morphological features, often appearing as flat or slightly elevated lesions with pale coloration, indistinct borders, and a mucus cap, predominantly in the proximal colon
27
. These characteristics contribute to higher miss rates and may partly explain the limited effectiveness of colonoscopy in preventing proximal colorectal cancer. It is estimated that 5% to 7% of colorectal cancers are classified as interval cancers, arising from lesions missed during screening colonoscopy despite adherence to recommended surveillance intervals
28
29
30
.

Several image-enhanced endoscopic techniques have been evaluated to improve SL detection
31
. Studies comparing narrow-band imaging with HD-WLE have not consistently demonstrated superior detection rates for SLs
32
33
34
. In contrast, AAC has been shown to enhance mucosal visualization by removing adherent mucus and inducing a transient acetowhite reaction, which increases surface contrast and facilitates lesion recognition
18
19
. Tribonias et al.
35
reported higher detection rates when acetic acid was used during a second inspection compared with saline, supporting a potential additive effect beyond reinspection alone.

Nevertheless, the sequential design of our study does not allow complete exclusion of a tandem observation effect. The act of reexamining the same colonic segment is known to increase lesion detection, particularly for conventional adenomas. However, prior evidence suggests that tandem inspection has a more limited impact on SLs compared with adenomas. The magnitude and lesion-specific nature of the increase observed in our study, restricted to proximal SLs, supports a biologically plausible additive role of acetic acid, rather than a nonspecific benefit of a second inspection alone.

Regarding clinical risk factors
36
37
, hypertension emerged as the only variable independently associated with proximal SL detection after multivariable adjustment. Although the CIs were relatively wide, reflecting a limited number of events, this finding is consistent with emerging evidence linking metabolic risk factors to serrated neoplasia
36
. These results should be interpreted cautiously and warrant confirmation in larger, adequately powered studies.

Some practical limitations and disadvantages of AAC should be acknowledged. Transient colonic spasm and increased mucus production have been described with acetic acid use. In our experience, these effects were infrequent and did not compromise mucosal evaluation. When necessary, additional suction, irrigation, or on-demand antispasmodic agents were sufficient to maintain adequate visualization.
This study has additional limitations. Colonoscopists were not blinded to acetic acid use, and the study was not randomized. Each patient served as their own control, which may introduce operator-related bias. Nonetheless, this design reflects routine clinical practice and allows assessment of the incremental value of acetic acid when applied during a systematic reinspection of the proximal colon.

Conclusions
In summary, this prospective study suggests that sequential use of 2% AAC during HD-WLE is associated with improved detection of proximal SLs. Although acknowledging the influence of reinspection, simplicity, low cost, safety profile, and ease of application of acetic acid supports its potential role as an adjunctive tool in daily colonoscopic practice, further randomized controlled trials with alternating inspection sequences are needed to isolate its independent effect on SSL detection.