Clinical value of rectal retroflexion during colonoscopy: a retrospective analysis

Abstract

Background & aims

Conventional forward-viewing colonoscopy may overlook lesions in the distal rectum, increasing the risk of missed colorectal neoplasia. Rectal retroflexion can improve visualization, but its diagnostic value remains debated. This study evaluated the diagnostic yield and safety of rectal retroflexion for detecting rectal polyps.

Methods

A total of 1,330 colonoscopy cases were retrospectively analyzed. Demographic characteristics, procedural success rate, and findings from both forward and retroflexed views were reviewed. The primary endpoints were the success rate of retroflexion, the per-patient polyp detection rate (PDR), and the per-patient polyp miss rate (PMR). Logistic regression was used to identify factors associated with retroflexion failure and additional polyp detection.

Results

Rectal retroflexion was successfully completed in 98.7% of cases, with no procedure-related adverse events. Rectal edema was the only independent predictor of retroflexion failure (OR = 54.8, P = 0.008). Retroflexion identified additional polyps in 76 patients (PMR = 5.7%), resulting in an overall PDR of 37.3% compared with 31.6% by forward view alone. Male sex (OR = 1.64), initial colonoscopy (vs. follow-up; OR = 2.62), and the presence of forward-view polyps (OR = 1.87) independently predicted additional polyp detection. Forward-view polyps > 5 mm were strongly associated with the detection of missed adenomas or sessile serrated adenomas (OR = 4.40).

Conclusions

Our findings suggest that rectal retroflexion is a safe and effective maneuver that can improve rectal polyp detection, particularly among male patients, those undergoing their first colonoscopy, and individuals with forward-view polyps, especially those > 5 mm. Routine inclusion of rectal retroflexion may enhance the quality and completeness of colonoscopic examinations.

Introduction

Colonoscopy is the gold standard for diagnosing colorectal diseases, offering direct visualization of the colonic mucosa and terminal ileum [1]. However, conventional forward-viewing colonoscopy faces limitations in examining the distal rectum and anal canal, an anatomically complex region characterized by rectal valves, angulation, and the anal sphincter [2, 3]. Consequently, lesions within these mucosal folds can be missed. To address this diagnostic challenge, the rectal retroflexion maneuver was developed [4]. By inverting the endoscope within the rectal vault, it provides a retroflexed view that allows for a more thorough mucosal inspection, theoretically enhancing the local diagnostic yield [5].

However, the routine implementation of rectal retroflexion is a subject of ongoing debate. Proponents argue that it enhances the detection of adenomas and early neoplasms, thus improving the quality of colonoscopy [6, 7]. In contrast, other studies have reported minimal impact on adenoma detection rate (ADR) , questioning its cost-effectiveness and necessity [8]. Compounding the controversy are concerns regarding its safety profile, including patient tolerance and the potential for severe complications like perforation [9–12].

Given these controversies, the present retrospective analysis of 1,330 consecutive colonoscopies was conducted to evaluate the diagnostic yield and safety of rectal retroflexion and to inform its potential role in routine endoscopic practice.

Methods

Study population

This retrospective study included consecutive patients who underwent colonoscopy performed by a single experienced endoscopist at the Department of Gastroenterology, Tianjin Union Medical Center, from January 2020 to December 2024. Findings from both forward and retroflexed rectal views were systematically recorded. Retroflexion was attempted in all patients unless contraindicated.

Exclusion criteria were as follows: (1) age < 18 years; (2) inadequate bowel preparation; (3) known rectal cancer or prior rectal surgery; (4) active inflammatory bowel disease (Crohn’s disease or ulcerative colitis); (5) colonoscopy performed for a planned therapeutic intervention in the rectum; (6) rectal polyps that could not be retrieved for histological analysis due to minute size or destruction during removal; and (7) incomplete medical records.

After applying these criteria, a total of 1,330 patients were included in the final analysis. The study was approved by the Institutional Review Board of Tianjin Union Medical Center (approval no. 2025-B165). The flow of patient inclusion and exclusion is summarized in Fig. 1.

Fig. 1.

The study flow diagram

Procedural methods

All examinations were performed with an Olympus CF-HQ290 series colonoscope. The retroflexion maneuver followed a standardized protocol. The scope was first advanced 10–15 cm from the anal verge. After adequate air insufflation of the rectal ampulla, the scope was gently advanced another 5–10 cm and then fully retroflexed by manipulating the control dials to achieve a 180°–210° bend. Successful retroflexion was confirmed by a clear view of the anorectal junction and the endoscope shaft. If needed, minor adjustments were made to obtain a complete circumferential view. The rectal mucosa was carefully inspected for polyps, hypertrophic anal papillae, hemorrhoids, and other lesions. All detected polyps were resected for histopathological assessment. The maneuver was deemed successful if a complete 360° view of the distal rectum was achieved.

The primary outcome measures were the per-patient polyp detection rate (PDR), the per-patient polyp miss rate (PMR) detected on retroflexion, and the retroflexion success rate. Two related metrics were used: (1) Forward-view PDR was defined as the proportion of patients with ≥ 1 polyp detected during forward view; (2) Overall PDR was defined as the proportion of patients with ≥ 1 polyp detected by combined forward and retroflexed views. PMR was defined as the proportion of patients with at least one additional polyp detected only on retroflexion. In accordance with the intention-to-treat principle, cases with failed retroflexion attempts were included in the denominator for analysis.

Data collection and verification

Demographic data (age and sex) were collected for all participants. Endoscopic findings, including polyp characteristics (number, size, location) and other mucosal abnormalities (e.g., hyperemia, edema, erosion), were documented. Histopathological results of all retrieved polyps were independently reviewed by two experienced pathologists.

Data were extracted from electronic endoscopy reports and archived static images by two independent gastroenterologists with over 5 years of colonoscopy experience, neither of whom performed the procedures. The reviewers, blinded to each other’s assessments, used a standardized form to independently verify polyp characteristics by cross-referencing written reports with corresponding images. Image view orientation (forward versus retroflexed) was determined by identifying anatomical landmarks before evaluating for polyps. For polyps detected during retroflexion, both reviewers confirmed the lesion was documented in retroflexed images, noted as such in the written report, and absent from forward-view images. A third senior gastroenterologist resolved any discrepancies to reach consensus.

Statistical analysis

Statistical analysis was performed using IBM SPSS Statistics for Windows, Version 23.0. Categorical variables were presented as counts and percentages and were compared using the chi-square test or Fisher’s exact test, as appropriate. Continuous variables were expressed as mean ± standard deviation (SD) and compared using the independent-samples t-test. Variables found to be significant in the univariate analysis were entered into a multivariate logistic regression model to identify independent predictors of polyp detection during retroflexion. All statistical tests were two-tailed, and a P-value < 0.05 was considered statistically significant.

Results

Patient characteristics

A total of 1,330 patients were included in the final analysis, with a mean age of 55.8 ± 11.5 years; 695 (52.3%) were male and 635 (47.7%) were female. The overall PDR for the entire cohort, based on the combination of forward-view and retroflexion findings, was 37.3% (496/1330). Other rectal findings such as hemorrhoids (83.2%) and hypertrophic anal papillae (9.9%) were common. The baseline demographic and clinical characteristics of the cohort are summarized in Table 1.

Table 1.

Baseline clinical characteristics of study subjects

Variable	All patients
	n = 1330 No. (%)
Age (years)	55.83 ± 11.46
Sex
Female	635(47.74)
Male	695(52.26)
Reason for colonoscopy
Initial colonoscopy	1129(84.89)
Follow-up colonoscopy	201(15.11)
Rectal hyperemia
Yes	16(1.20)
No	1314(98.80)
Rectal bleeding
Yes	8(0.60)
No	1322(99.40)
Rectal edema
Yes	6(0.45)
No	1324(99.55)
Rectal erosion
Yes	7(0.53)
No	1323(99.47)
Rectal scarring
Yes	30(2.26)
No	1300(97.74)
Any polyp found in the rectum
On forward-view examination	420(31.58)
On retroflexed examination	76(5.71)
Findings of Rectal Retroflexion Examination
Normal	177(13.31)
Polyps	76(5.71)
Hemorrhoids	1107(83.23)
Hypertrophic Anal Papillae	132(9.92)
Anal Papilloma	11(0.83)

Procedural outcomes and safety

Rectal retroflexion was successfully performed in 1,313 of 1,330 cases, achieving a success rate of 98.7% (Fig. 2). The 17 (1.3%) failures were attributed to factors such as rectal stenosis or a shortened rectum. In the initial screening for risk factors, rectal hyperemia, edema, erosions, and bleeding were associated with retroflexion failure (all P < 0.05) (Tables 2 and 3). However, in the multivariate logistic regression, only rectal edema was identified as an independent risk factor for failure (OR = 54.80, 95% CI: 2.90–1036.70; P = 0.008) (Table 3). The wide confidence interval suggests this estimate should be interpreted with caution, likely due to the small number of events. Regarding safety, no immediate procedure-related adverse events, such as perforation, significant bleeding, or mucosal tears, were observed in any patient. Long-term adverse events were not monitored in this study.

Fig. 2.

Retroflexion in the rectum

Table 2.

Characteristics of patients in whom retroflexion in the rectum was unsuccessful

Variable	Successful retroflexion n = 1313 No.(%)	Unsuccessful retroflexion n = 17 No.(%)	P value
Age (years)	55.81 ± 11.456	56.71 ± 11.920	0.535
Sex			0.585
Female	628(47.83)	7(41.18)
Male	685(52.17)	10(58.82)
Rectal hyperemia			0.017*
No	1299(98.93)	15(88.24)
Yes	14(1.07)	2(11.76)
Rectal bleeding			0.098
No	1306(99.47)	16(94.12)
Yes	7(0.53)	1(5.88)
Rectal edema			0.002**
No	1309(99.70)	15(88.24)
Yes	4(0.30)	2(11.76)
Rectal erosion			0.086
No	1307(99.54)	16(94.12)
Yes	6(0.46)	1(5.88)
Rectal scarring			0.323
No	1284(97.79)	16(94.12)
Yes	29(2.21)	1(5.88)

*p < 0.05

**p < 0.01

Table 3.

Univariate and multivariate logistic analysis of risk factors associated with unsuccessful rectum retroflexion

Variable	Univariate Analysis		P	Multivariate Analysis		P
	OR	95% CI		OR	(95% CI)
Age (years)	1.007	0.965–1.050	0.750
Sex
Female	1.000
Male	1.310	0.496–3.462	0.586
Rectal hyperemia
No	1.000
Yes	12.371	2.583–59.262	0.002**	3.690	1.599–14.315	0.301
Rectal bleeding
No	1.000
Yes	11.661	1.355-100.357	0.025*	2.593	0.064-104.255	0.613
Rectal edema
No	1.000
Yes	87.400	11.537-662.129	<0.001***	54.799	2.895-1036.696	0.008**
Rectal erosion
No	1.000
Yes	20.453	2.164-193.294	0.008**	0.445	0.006–31.781	0.710
Rectal scarring
No	1.000
Yes	2.767	0.355–21.572	0.331

*p < 0.05

**p < 0.01

***p < 0.001

Polyp detection rates

During the initial forward-view examination, rectal polyps were detected in 420 patients, yielding a forward-view PDR of 31.6% (420/1330). This included a baseline ADR of 9.0% (119/1330) and a sessile serrated adenoma (SSA) detection rate of 3.4% (45/1330). Subsequent retroflexion identified additional polyps in 76 patients, corresponding to a PMR of 5.7% (76/1330). Importantly, this technique found an additional 21 clinically significant lesions (14 adenomas and 7 SSAs), resulting in a combined miss rate for adenomas and SSAs of 1.6% (21/1330).

Combining both viewing modalities, the overall PDR for the entire cohort was 37.3% (496/1330), representing a 5.7% point increase (18.1% relative increase) attributable to retroflexion. The characteristics of polyps detected during forward and retroflexed views, including their size, number, and histology, are detailed in Table 4. Notably, the majority of missed polyps (94.7%) were ≤ 5 mm in size.

Table 4.

Comparison of forward view and retroflexed view in rectal

Variable	All patients (n = 1330)
	Forward view	Retroflexed view
Polyp detection rate (per-patient), %	31.58(420)	5.71(76)
Number of polyps per patient, No. (%)
1	151(35.95)	50(65.79)
2	42(10.00)	2(2.63)
≥ 3	227(54.05)	24(31.58)
Size of polyps per patient, No. (%)
≤ 5	358(85.24)	72(94.74)
6–9	35(8.33)	2(2.63)
≥ 10	27(6.43)	2(2.63)
Histology, No. (%)
Inflammation	40(9.52)	10(13.16)
Hyperplastic	210(50.00)	45(59.21)
Adenoma	119(28.33)	14(18.42)
Sessile serrated adenoma	45(10.71)	7(9.21)
Adenocarcinoma	6(1.43)	0(0.00)

Risk factors for polyp detection on retroflexion

On bivariate analysis, male sex, undergoing an initial colonoscopy, and the presence of polyps on forward view were all significantly associated with the detection of additional polyps during retroflexion (Tables 5 and 6). The analysis confirmed that male sex (OR = 1.64, 95% CI: 1.00–2.67; P = 0.048), initial colonoscopy (OR = 2.62, 95% CI: 1.04–6.60; P = 0.041), and the presence of polyps on forward view (OR = 1.87, 95% CI: 1.17–3.00; P = 0.009) were independent predictors for detecting additional polyps during retroflexion (Table 6).

Table 5.

Characteristics of patients in whom polyps were identified in retroflexion

Variable	Presence of polyps in retroflexion n = 76 No. (%)	No polyps in retroflexion n = 1254 No.(%)	P value
Age (years)	55.75 ± 9.628	55.83 ± 11.563	0.973
Sex			0.028*
Female	27(35.53)	608(48.48)
Male	49(64.47)	646(51.52)
Reason for colonoscopy			0.032*
Initial colonoscopy	71(93.42)	1058(84.37)
Follow-up colonoscopy	5(6.58)	196(15.63)
Presence of rectal polyps on forward view
No	40(52.63)	870(69.38)	0.002**
Yes	36(47.37)	384(30.62)
Diameter of rectal polyps on forward view(mm)
≤ 5	70(92.11)	1198(95.53)
>5	6(7.89)	56(4.47)	0.161

*p < 0.05

**p < 0.01

Table 6.

Univariate and multivariate logistic analysis of risk factors associated with detecting polyps in retroflexion

Variable	Univariate Analysis		P	Multivariate Analysis		P
	OR	95% CI		OR	95% CI
Age(years)
18–44	1.000
45–59	1.009	0.500-2.039	0.979
60–74	1.518	0.749–3.074	0.246
75	0.261	0.033–2.059	0.204
Sex
Female	1.000
Male	1.708	1.054–2.768	0.030*	1.636	1.004–2.667	0.048*
Reason for colonoscopy
Follow-up colonoscopy	1.000
Initial colonoscopy	2.632	1.049–6.579	0.039*	2.625	1.042–6.579	0.041*
Presence of rectal polyps on forward view
No	1.000
Yes	2.039	1.280–3.249	0.003**	1.871	1.168–2.998	0.009**
Diameter of rectal polyps on forward view (mm)
≤ 5	1.000
>5	1.834	0.764–4.402	0.175

*p < 0.05, **p < 0.01

Risk factors for detecting clinically significant polyps

Given that rectal retroflexion identified an additional 21 precancerous lesions (adenomas and SSAs) that were initially missed, a sub-analysis was conducted to identify predictors specifically for missed adenomas or SSAs. The multivariate model revealed that the presence of a forward-view polyp > 5 mm was the sole independent risk factor for detecting these clinically significant lesions on retroflexion (OR = 4.40, 95% CI: 1.35–14.33; P = 0.014), as shown in Table 7.

Table 7.

Univariate and multivariate logistic analysis of risk factors associated with detecting adenomas and SSAs in retroflexion

Variable	Presence of adenoma/ SSAs in retroflexion n = 21 No.(%)	No adenoma/ SSAs in retroflexion n = 1309 No.(%)	Univariate Analysis		P	Multivariate Analysis		P
			OR	95% CI		OR	95% CI
Age (years)	56.62 ± 11.838	55.81 ± 11.456
18–44	3(14.29)	212(16.20)	1.000
45–59	7(33.33)	613(46.83)	0.807	0.207–3.149	0.758
60–74	10(47.62)	413(31.55)	1.711	0.466–6.283	0.418
75	1(4.76)	71(5.42)	0.995	0.102–9.722	0.997
Sex
Female	9(42.86)	626(47.82)	1.000
Male	12(57.14)	683(52.18)	1.222	0.511–2.920	0.652
Reason for colonoscopy
Initial colonoscopy	19(90.48)	1110(84.80)	1.000
Follow-up colonoscopy	2(9.52)	199(15.20)	0.587	0.136–2.540	0.476
Presence of rectal polyps on forward view
No	9(42.86)	901(68.83)	1.000
Yes	12(57.14)	408(31.17)	2.944	1.231–7.043	0.015*	1.997	0.738–5.402	0.173
Diameter of rectal polyps on forward view(mm)
≤ 5	16(76.19)	1252(95.65)	1.000
>5	5(23.81)	57(4.35)	6.864	2.429–19.394	<0.001 ***	4.398	1.350-14.333	0.014 *

*p < 0.05

***p < 0.001

Discussion

In this retrospective analysis, we observed that rectal retroflexion was a feasible and safe maneuver that provided incremental detection of rectal polyps. It identified additional polyps in 76 patients who were otherwise considered clear on forward view. This finding highlights the potential value of retroflexion in complementing conventional forward-viewing colonoscopy and suggests that this technique may be considered for inclusion in standard colonoscopy protocols, particularly in certain patient subgroups.

Our study indicates that rectal retroflexion is a feasible procedure in most patients, with a success rate of 98.7%, which is higher than the 93.9% reported in a previous large prospective study [13]. Importantly, no procedure-related adverse events were observed, reinforcing the technique’s favorable safety profile for diagnostic purposes. Our analysis identified rectal edema as a significant independent risk factor for retroflexion failure (OR = 54.80; Table 3). While the wide confidence interval warrants caution, this finding is biologically plausible, as edema can reduce rectal compliance and lumen space. This suggests that endoscopists should exercise careful judgment when considering retroflexion in patients with evident rectal inflammation.

The primary contribution of our study lies in the incremental diagnostic yield of retroflexion. The overall PDR of 37.3% in our cohort is consistent with established benchmarks for average-risk screening populations, confirming the baseline quality of our examinations [14]. The reported 9.0% ADR must be interpreted within the specific context of this study. While established quality benchmarks typically recommend an ADR of ≥ 25% [15], it is crucial to note that these standards reflect whole-colon findings. In contrast, our data strictly represents rectal ADR. The prevalence of adenomas in the distal rectum is inherently lower than in the proximal colon, whereas hyperplastic polyps are more frequent in this region [16–18]. This anatomical distinction is mirrored in our high overall PDR of 37.3%, which confirms effective lesion detection but highlights the predominance of non-neoplastic histology in the rectum. Furthermore, our cohort’s younger mean age (55.8 years) and strict exclusion criteria likely contribute to a lower adenoma prevalence compared to older screening populations.

Despite this lower baseline prevalence of adenomas, the utility of retroflexion remained significant. We observed that the 5.7% PMR (including a combined 1.6% miss rate for adenomas and SSAs; Table 4) appears substantially higher than that reported in many earlier key studies. For instance, several studies by Cutler, Grobe, and Saad reported adenoma miss rates of less than 1% or even zero with retroflexion [8, 13, 19]. The higher detection rates in our cohort may be attributed to several factors. First, all procedures were performed by a single, highly experienced endoscopist, whose meticulous technique may have enhanced detection. Second, our study comprised a higher proportion of male participants, which likely contributed to the elevated detection rates through multiple mechanisms. Males consistently demonstrate higher rates of colorectal adenomas and cancer compared to females [20, 21], resulting in greater baseline polyp burden and consequently more opportunities for detection during retroflexion. This sex disparity is multifactorial, influenced by sex-related lifestyle factors such as diet, smoking, and alcohol consumption, all of which affect both polyp burden and distribution patterns. Additionally, the protective effect of estrogen against colorectal neoplasia in females may further accentuate the difference in polyp prevalence between sexes [22, 23]. The interplay of these biological and behavioral factors underscores the particular clinical value of routine retroflexion in male patients, who may derive greater incremental benefit from this supplementary examination. Finally, characteristics specific to our study population, including potential regional dietary patterns and lifestyle factors prevalent in China, may contribute to an elevated baseline prevalence of rectal neoplasia in our cohort. This highlights that the value of retroflexion may be population-dependent and more pronounced in high-prevalence settings.

A key practical outcome of our study is the identification of patient subgroups who derive the greatest benefit from retroflexion. Our multivariate analysis confirmed that male sex, initial (first-time) colonoscopy, and the presence of polyps on forward view were independent predictors for detecting missed polyps (Table 6). Notably, a forward-view polyp > 5 mm was the sole independent predictor for detecting a missed adenoma or SSA (OR = 4.40; Table 7). This novel and clinically actionable finding suggest that the presence of a larger polyp on forward view should be considered a key indicator for a heightened risk of synchronous, clinically significant lesions nearby. While previous studies have explored risk factors for polyp detection during retroflexion in the right colon [24, 25], to our knowledge, this is the first study to systematically analyze these risk factors specifically for the rectal region.

This study has several limitations. First, all procedures were performed by a single experienced endoscopist at our center. This design reduced inter-operator variability, but limited generalizability. Our findings largely reflected this operator’s skill and practice style. Endoscopists with different training, experience, or technique may achieve different detection and complication rates. These results should be cautiously applied to other settings, and multi-center studies with multiple operators are needed for validation. Second, our study focused solely on diagnostic yield. We did not assess therapeutic applications, such as retroflexed polypectomy, nor did we compare retroflexion to anoscopy for non-neoplastic conditions like hemorrhoids. Third, while we observed no complications, this reflected the safety when performed by an experienced operator and may not apply to all practitioners. Less experienced endoscopists may face higher complication risks. Our data showed the technique’s safety in expert hands, not necessarily across all settings. Adequate training and supervised skill development are essential before routine use. Studies including endoscopists with varying experience levels would better characterize the safety profile. Future large-scale, multi-center prospective studies are needed to validate our findings.

In conclusion, rectal retroflexion is a safe and effective technique for enhancing the detection of rectal polyps, especially for clinically significant adenomas and SSAs that may be missed by conventional forward view. Based on our findings, we recommend that rectal retroflexion may be routinely considered, particularly for male patients, those undergoing their first colonoscopy, and individuals with polyps already detected on forward view. For patients with forward-view polyps > 5 mm, retroflexion should be considered to enhance the quality of endoscopic examination.

Authors’ contributions

Q.J.H. and S.Y.Z. conceived and designed the project. Q.J.H., W.L., Y.W.W., Y.R.L., Y.Y.L., and Q.Z. analyzed and interpreted the data. Q.J.H., W.L., and Y.W.W. drafted the manuscript. S.Y.Z. and W.L. critically revised the article for important intellectual content. All authors read and approved the final manuscript.

Funding

This work was supported by grants from National Key Research and Development Program of China (No. 2022YFC3602105) and a Hospital-level Project of Tianjin Union Medical Center (No. 2021YJ018).

Data availability

All data used in the study are presented in the tables.

Declarations

Ethics approval and consent to participate

The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the Ethical Committee of Tianjin Union Medical Center (approval number: 2025-B165) and individual consent for this retrospective analysis was waived.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.