Colonoscopy surveillance after polypectomy
- First Online:
- Cite this article as:
- Kasugai, K., Ogasawara, N. & Sasaki, M. Clin J Gastroenterol (2011) 4: 355. doi:10.1007/s12328-011-0246-5
- 56 Views
The goal of surveillance examinations after polypectomy is to detect new adenomas and missed synchronous adenomas, as well as preventing adenomas from becoming invasive or cancerous. The first colonoscopy surveillance program reported was the National Polyp Study from the United States in 1997, with an update in 2003. First screening colonoscopy and polypectomy have been shown to produce the greatest effects in reducing the incidence of colorectal cancer in patients with adenomatous polyps. However, a large number of adenomas are being discovered as a result of the increased use of colorectal cancer screening, particularly with the dramatic increase in screening colonoscopy and surveillance. Increased efficiency of surveillance colonoscopy practices is therefore needed to decrease the cost, risk, and overuse of medical resources. In developing surveillance programs, studying miss rates and incidences and performing separate evaluations are important, along with accurately assessing incidence. This is because the recurrence rate or apparent incidence after colonoscopic polypectomy includes the true incidence of new polyp formation plus the incidence of missed polyps from the initial colonoscopy. Many studies have indicated the number of adenomas on initial examination as the most significant predictor for missed adenoma and incidence of adenoma on surveillance colonoscopy. In Japan, many facets of colonoscopic examination differ from those in Western countries. Further studies are recommended to establish an appropriate and original Japanese colonoscopy surveillance program for use after polypectomy, based on guidelines from the United States.
According to reports from the Japanese Ministry of Health, Labour and Welfare, colorectal cancer (CRC) was the third leading cause of cancer death in Japan in 2009, and the leading cause for women. The incidence of CRC had steadily increased in Japan over the past 20 years. In the United States, CRC is the second most common cause of cancer death, and incidence has been falling since the mid-1980s. This decline in CRC incidence in the United States is commonly attributed to CRC screening . In contrast to the United States, the westernization of lifestyle in Japan, particularly in terms of dietary habits, has been marked since 1950 and is presumed to be directly involved with the rising incidence of CRC . These increases in CRC incidence in Japan have occurred despite public policies endorsing CRC screening. CRC prevention has thus become an important problem in Japan.
Adenomatous polyps are the most common neoplasm identified during screening colonoscopy [3, 4] and are typically removed by colonoscopic polypectomy to reduce the risk of developing into CRC or advanced adenoma [5–8]. Furthermore, patients with adenomas are usually entered into a periodic colonoscopic surveillance program to detect missed synchronous lesions, early local persistence or recurrences and multiple new metachronous lesions [9, 10].
Until the early 1990s, no surveillance guidelines were available to address how clinicians should follow-up post-polypectomy patients. The common practice was for these patients to undergo annual follow-up colonoscopy to detect new additional adenomas and missed synchronous adenomas . In the United States, the National Polyp Study reported in 1993 , involved a randomized trial comparing 1- and 3-year surveillance intervals after polypectomy; surveillance at 3-year intervals has since become standard practice for the majority of patients after polypectomy . However, recommendations for post-polypectomy surveillance in Japan have not been established. In current practice, follow-up intervals between colonoscopies after polypectomy are variable; although often annual, they are not based on data from randomized clinical trials . Colonoscopy in Japan differs from that in other countries, with methods such as bowel preparation using polyethylene glycol (PEG) solution given in the morning of the day of colonoscopy, examination without sedative drugs, and regular use of dye-spray chromoscopic examination for detection of diminutive or non-polypoid neoplasia [14–16].
In this review we assess miss rates of colonoscopy, incidence of neoplasm post-polypectomy and surveillance programs after polypectomy for individuals at average risk, but not for individuals with a personal or family history of CRC or adenomas, inflammatory bowel disease, or high-risk genetic syndromes. Finally, we mention adaptive surveillance guidelines for Japanese patients after polypectomy.
The incidence of colorectal polyps at surveillance colonoscopy has been reported in several studies [12, 17–20]. However, whether the polyps found at surveillance colonoscopy are true metachronous polyps or are missed synchronous polyps remains unclear. The recurrence rate or apparent incidence after colonoscopic polypectomy includes the true incidence of new polyp formation along with that of missed polyps at the initial colonoscopy . Currently, colonoscopy is the best available method to detect and remove colonic polyps and can therefore be considered as the gold standard for this purpose. However, colonoscopy is not infallible and most gastroenterologists are aware that polyps can be missed during colonoscopy.
The most reliable method to investigate the polyp miss rate is ‘tandem or ‘back-to-back’ colonoscopy. This method involves performing two consecutive same-day colonoscopies for each patient. In such studies, the miss rate is usually expressed as the number of polyps detected only during the second colonoscopy relative to the total number of polyps found during both examinations. The first study of colonoscopy miss rates by tandem colonoscopy was reported by Hixson et al. . In 90 patients, 24 neoplastic polyps were missed. Miss rates were 16% for neoplastic polyps ≤5 mm, 12.3% for polyps 6–9 mm and 0% for polyps ≥1 cm. A larger tandem colonoscopy study was reported by Rex et al. . In this study, for 183 randomized patients with same-day back-to-back colonoscopy, the miss rate was 6% for adenomas ≥1 cm, 13% for adenomas 6–9 mm, and 27% for adenomas ≤5 mm. The overall miss rate for adenomas was 24%. Percentages of patients with 0, 1, or >1 adenomas on first examination who had ≥1 adenomas on second colonoscopy were 16%, 30%, and 48%, respectively. The overall miss rate for patients was 30%. Another method of studying miss rates using a transparent cap for colonoscopy has been reported . In 24 patients with colorectal polyps, tandem colonoscopic procedures were carried out on the same day in random order, first without a transparent cap and then with the cap (without-to-with), or first with the cap and then without the cap (with-to-without). The miss rate for polyps was 15% in without-to-with tandem colonoscopies, and 0% in with-to-without procedures (P = 0.0125). Harrison et al.  reported miss rates in 98 patients who were randomized to a second examination of the proximal colon using either a forward or retroflexed view. The calculated miss rate for adenomas among patients with the forward view was 33.3%, compared to 23.7% with the retroflexed view; however, there was no difference in miss rates for adenomas (P = 0.31) between these two different views for the second examination.
A systematic review evaluating only miss rates by tandem colonoscopy covered six studies with a total of 465 patients . The pooled miss rate tended to be higher for non-adenomatous polyps (27%) than for adenomas (22%), but this difference was not significant. (P = 0.49). Miss rates according to size were 26% for adenomas ≤5 mm, 13% for adenomas 5–10 mm, and 2.1% for adenomas ≥10 mm. These miss rates increased with decreasing polyp size and larger polyps are, unsurprisingly, less likely to be missed. The most recent back-to-back study was reported by Heresbach et al. , and was performed as a large multicenter study, with same-day back-to-back video colonoscopy for 286 patients from 11 centers. Miss rates for polyps, adenomas, polyps ≥5 mm, adenomas ≥5 mm, and advanced adenomas were 28%, 20%, 12%, 9% and 11%, respectively. However, these study designs might involve differences from routine clinical practice, and colonoscopists in studies using tandem or back-to-back colonoscopy would be aware that the patients were participating in a miss-rate study and would thus be more likely to perform second examinations in a most meticulous manner. Miss rates in such studies might therefore be higher than in clinical practice. Other studies determining colonoscopic miss rates in routine clinical practice have been reported. In 76 patients with a mean of 47 days between colonoscopic examinations, the miss rate for neoplastic polyps was 12% . A total of 17% of patients had ≥1 neoplastic polyp missed on initial examination. Another large study involved 688 patients who underwent total colonoscopy twice within 30 days . In 157 of 688 patients (22.8%), 200 neoplastic lesions (all adenomas) were missed during the first examination. The pooled miss rate for adenoma was 14.9% and miss rates according to size were 15.9% for adenomas ≤5 mm, 11.9% for adenomas 5–10 mm and 2.3% for adenomas >10 mm. The different miss rates in these studies might have been attributable to the meticulous dye-spraying chromoscopic examination or the subject population size. Chromoendoscopy reportedly improves the detection of small adenomas and facilitates the visualization of mucosal lesions  and could reduce or eliminate colorectal neoplasia miss rates during colonoscopy . Several other new developments that improve colonoscopic polyp detection are already available or currently under investigation. Future advances in image-enhanced endoscopy [30, 31], such as narrow-band imaging [32–34], autofluorescence imaging [35–37], and Fujinon Intelligent Color Enhancement  have been developed and these new technologies may improve the ability of polyp detection and reduce miss rates. Although mounting evidence suggests that these new technologies are superior to conventional endoscopy, current guidelines are limited. Further large-scale randomized controlled trials comparing these modalities in different patient subpopulations are warranted .
A recent study by Kim et al.  investigated 285 patients undergoing screening colonoscopy followed within 2 months by colonoscopy for polypectomy to determine miss rates for polyps and conditions under which missed polyps on colonoscopy could be reduced. Of 1326 polyps identified, 350 were initially missed (26.4%). Patient-related factors included male sex [odds ratio (OR) 2.11, 95% confidence interval (CI) 1.21–3.70) and >60 years (OR 2.51, 95% CI 1.48–4.30). Polyp-related factors included >4 polyps (OR 4.48, 95% CI 1.91–10.5). However, large polyps were less frequently missed on colonoscopy (OR 0.10, 95% CI 0.05–0.19).
Miss rates for neoplastic polyps
Number of patients
Miss rates per patients (%)
Neoplasm miss rates
Over all (%)
Hixso et al. 
≤5 mm 16%, 6–9 mm 12.3%, ≥10 mm 0%
Rex et al. 
≤5 mm 27%, 6–9 mm 13%, ≥10 mm 6%
Matsushita et al. 
Tandem colonoscopy with or without a transparent cap
≤5 mm 6.6%, 6–9 mm 1.1%, ≥10 mm 0%
Bensen et al. 
Within 120 days between colonoscopy
≤5 mm 15%, 6–9 mm 13%, ≥10 mm 7%
Rex et al. 
Tandem colonoscopy, wide angle and standard
≤5 mm 25%, 6–9 mm 0%, ≥10 mm 0%
Harrison et al. 
Tandem colonoscopy by forward view or a retroflexed view
≤5 mm 27.5%, 6–9 mm 1.3% ≥ 10 mm 0%
Kasugai et al. 
Within 30 days between colonoscopy
≤5 mm 16%, 6–9 mm 12%, ≥10 mm 2%
Heresbach et al. 
Tandem or back-to-back colonoscopy
<5 mm 27%, ≥5 mm 9%
Kim et al. 
Within 2 months between colonoscopy
<5 mm 11%, 5–9 mm 14%, ≥10 mm 2%a
Miss rates for adenomas and cancers differ between examiners. Gastroenterologists in general have a miss rate for CRC of approximately 3%, compared to around 13% for non-gastroenterologists . Two experienced colonoscopists in one study showed differing miss rates for adenomas (17 vs. 48%), subsequently shown to be correlated to differences in colonoscopic technique . Methods of withdrawal vary between colonoscopists and may represent a major factor in miss rates [21, 53]. Further evaluation of miss rates in terms of procedure-related factors identified factors of colonoscopy by clinical fellows (OR 2.20, 95% CI 1.02–4.84) and delayed insertion time (OR 4.10, 95% CI 2.14–7.86) .
Another factor that may have influenced polyp detection was suboptimal bowel preparation. Lebwohl et al.  reported adenoma miss rates for patients who underwent initial examinations with suboptimal bowel preparations and repeat examination with optimal preparation. Miss rates from colonoscopies repeated in <3 years for adenoma and advanced adenoma were 42% and 27%, respectively. For colonoscopies repeated in <1 year, adenoma and advanced adenoma miss rates were 35% and 36%, respectively. These high miss rates for colonoscopies performed with suboptimal bowel preparation might warrant early follow-up examination.
The number of neoplastic polyps in the initial examination is a significant predictor of missed neoplastic polyps . Miss rates according to the number of polyps found on first examination were 5.3% for none, 19.6% for one, 23.5% for two and 35.5% for ≥3 polyps detected on first examination. The OR for finding adenomas on the second examination was higher for patients with ≥2 adenomas on first examination (OR 2.06, 95% CI 1.42–3.00) than for those of 0 or 1 adenomas on first examination . Other studies demonstrated similar findings with patients showing 1, 2 or ≥3 detected adenomas on initial colonoscopy, with miss rates of 17%, 29% and 42%, respectively, in one report , and 30%, 48% and 47%, respectively, in a second report . These data remind us that colonoscopists should continue to perform meticulous examinations with the aim of detecting other polyps when multiple polyps are detected.
Incidence of colon polyps refers to the rate at which individuals develop colon polyps over a specified interval of time. The concept is similar to the recurrence rate in post-polypectomy surveillance studies of patients who have previously been found to have a ‘clean colon’ by colonoscopy . The recurrence rate or apparent incidence after colonoscopic polypectomy actually includes the true incidence of new polyp formation and that of polyps missed on initial colonoscopy. Studying miss rates and incidence rates and performing separate evaluations is thus important to accurately assess incidences.
Patients previously shown to have adenomas may have an increased biological potential to develop adenomas, so the incidence of new polyps may be higher in these patients than in patients without previous adenomas . In addition, these patients have a greater risk of missed adenomas on initial examination [20, 23, 28, 55]. The cumulative incidence in patients who showed no neoplastic polyps at index colonoscopy was 14.5% at 3 years and 27.3% at 5 years . Neugut et al.  reported a 36-month cumulative incidence of 16% for adenoma in patients with no abnormalities at index colonoscopy. In addition, Rex et al.  reported a 5-year incidence of 27% for adenoma after negative colonoscopy. In contrast, incidences of adenoma at an interval of 6 months to 4 years among patients with one or more neoplastic polyps at index colonoscopy were 30–50% [12, 18, 58–60]. A higher incidence of adenomas in post-polypectomy studies has been associated with polyp multiplicity at index colonoscopy [12, 28, 61], increased size of index adenomas , older age [12, 28], and family history of a parent with CRC .
Most incident adenomas are small, and the mean size of incident adenomas is substantially smaller than that of prevalent adenomas [19, 21]. A substantial proportion of small neoplastic polyps detected by post-polypectomy surveillance colonoscopy were missed at the initial colonoscopy [23, 28]. Bensen et al.  determined colonoscopic miss rates in routine clinical practice and true 1-year recurrence rates of colorectal neoplastic polyps in 79 patients with a mean of 47 days (range 1–119 days) between colonoscopic examinations. A total of 17% of patients had ≥1 neoplastic polyps missed at the initial examination. The observed 1-year recurrence rate was 28%, and by comparing this to the proportion of patients with ≥1 missed neoplastic polyp, the true 1-year recurrence of neoplastic polyps was calculated as 11%. Another study indicated miss rates for patients with 1, 2, or ≥3 neoplastic polyps on first examination were 19.6%, 23.5%, and 35.5%, respectively, while 1-year cumulative incidence rates for patients with 1, 2, or ≥3 neoplastic polyps on first examination were 18.4%, 21.1%, and 34.2%, respectively. Miss rates and 1-year cumulative incidence rates seem quite similar. This numerical correlation might indicate that most neoplastic polyps detected at post-polypectomy surveillance colonoscopy within 1 year were missed at the original colonoscopy . These studies evaluated either miss rates or incidence regardless of the number of polyps detected at index colonoscopy. This is a very important factor, because the number of neoplastic polyps on initial examination is a significant predictor of the risk of missed neoplastic polyps [20, 23, 28, 55] and the incidence of neoplastic polyps on surveillance colonoscopy [12, 61].
The rationale for colonoscopic surveillance has always been based on the high detection rate for colorectal adenomas on follow-up (30–50%) after complete clearance colonoscopy [12, 18, 58–60]. Adenomas with advanced pathology have a much higher malignant potential  and the object of colonoscopic surveillance is to ensure that such lesions are detected before becoming invasive. In fact, patients who have undergone polypectomy and long-term surveillance have been shown to have a reduced incidence of CRC [5–8]; however, subgroups can be identified as having higher or lower risk of subsequent cancer. Supporting the concept of stratifying patients by baseline factors is thus important, so that patients at increased risk can be identified for more intensive surveillance and patients at lower risk can be identified for less intensive surveillance.
Surveillance recommendations 
Patients with small rectal hyperplastic polyps should be considered to have normal colonoscopies, and therefore the interval before the subsequent colonoscopy should be 10 years; an exception is patients with hyperplastic polyposis syndrome who are at increased risk for adenomas and colorectal cancer and need to be identified for more intensive follow-up evaluation
Patients with only 1 or 2 small (<1 cm) tubular adenomas with only low-grade dysplasia should have their next follow-up colonoscopy in 5–10 years; the precise timing within this interval should be based on other clinical factors (such as prior colonoscopy findings, family history, and the preferences of the patient and judgment of the physician)
Patients with 3–10 adenomas, or any adenoma ≥1 cm, or any adenoma with villous features, or high-grade dysplasia should have their next follow-up colonoscopy in 3 years providing that piecemeal removal has not been performed and the adenoma(s) are removed completely; if the follow-up colonoscopy is normal or shows only 1 or 2 small tubular adenomas with low-grade dysplasia, then the interval for the subsequent examination should be 5 years
Patients who have >10 adenomas at 1 examination should be examined at a shorter (<3 years) interval, established by clinical judgment, and the clinician should consider the possibility of an underlying familial syndrome
Patients with sessile adenomas that are removed piecemeal should be considered for follow-up evaluation at short intervals (2–6 months) to verify complete removal; once complete removal has been established, subsequent surveillance needs to be individualized based on the endoscopist’s judgment. Completeness of removal should be based on both endoscopic and pathologic assessments
More intensive surveillance is indicated when the family history may indicate heriditary non-polyposis CRC
Polyp and patient characteristics at baseline colonoscopy have been identified in a number of studies and have been considered to potentially stratify patients according to the risk of advanced adenoma at subsequent colonoscopy. Predictors of advanced adenoma or cancer include ≥3 adenomas, villous elements in any adenoma, any adenoma with high-grade dysplasia, and any adenoma ≥1 cm in size [10, 65, 66].
Multiplicity at index colonoscopy has been shown to predict subsequent detection of advanced adenomas . Atkin et al.  followed a cohort of patients who underwent initial removal of rectosigmoid adenomas without further intervention in the colon for an average of 13.8 years. They showed that having ≥2 rectosigmoid adenomas compared with 1 rectosigmoid adenoma at baseline was associated with an increased risk for subsequent colon cancer, but not for subsequent rectal cancer. Another cohort study indicated that the number of adenomas at baseline was related to an increased risk (OR 1.25, 95% CI 1.13–1.38) of advanced adenoma at surveillance .
Adenoma size >1 cm has also been shown to predict metachronous advanced adenoma in some studies [6, 62, 67]. Yang et al.  showed that larger adenoma size was related to subsequent risk of advanced neoplasia at surveillance. The relative risk is higher for polyps ≥1.0 cm compared with polyps <0.5 cm (OR 4.4, 95% CI 1.9–10.2). Histologic type of adenoma at baseline was an important predictor of advanced neoplasia or CRC. Atkin et al.  showed that tubulovillous histology at baseline was associated with a standardized incidence ratio (SIR) of 3.8 (95% CI 2.2–6.0) and villous histology had an SIR of 5.0 (95% CI 2.2–9.9) for the detection of subsequent colon cancer. Yang et al.  reported that villous or tubulovillous histology at baseline indicated a higher risk (OR 8.34, 95% CI 3–16.0) for the detection of advanced neoplasms (rectal cancer or adenoma with severe dysplasia) at follow-up colonoscopy. Furthermore, high-grade dysplasia is related to larger adenoma size and villous component at baseline and is an important predictor for subsequent advanced neoplasia [6, 68].
From these many studies, a consensus can be seen that the group at lower risk for subsequent advanced adenomas was defined as only 1 or 2 adenomas, <1 cm in size with no high-grade dysplasia or villous features. The risk for colon cancer in such low-risk patients could be considered to be similar to that in the average-risk population . Therefore, in the surveillance guideline (Table 2), lower-risk group patients were recommended to have a follow-up colonoscopy in 5–10 years, whereas individuals with hyperplastic polyps only needed a 10-year follow-up evaluation, as for the average-risk population .
However, recommendations for post-polypectomy surveillance have not been established in Japan. In current practice, intervals between colonoscopies after polypectomy are variable, often annual, and not based on data from randomized clinical trials. The first large multicenter retrospective cohort study to analyze the incidence of advanced neoplasia after initial colonoscopy in Japan has been reported . A total of 5309 patients were enrolled and divided into four groups based on the initial colonoscopy: A, patients without any adenoma; B, with only adenomas <6 mm; C, with adenomas of ≥6 mm; or D, with any intramucosal cancer. A total of 379 index lesions (large adenoma ≥10 mm and intramucosal/invasive cancer) were newly diagnosed during follow-up colonoscopy and the cumulative incidence of index lesions in each group were 2.6%, 6.7%, 13.4%, and 12.6%, respectively. Moreover, the cumulative incidence of index lesions at 1 and 3 years was 0.5/1.9% and 2.7/5.6% in Groups A + B (low-risk group) and Groups C + D (high-risk group), respectively. According to this evidence, surveillance colonoscopy after initial total colonoscopy was recommended after 3 years for patients without any polyps or with polyps <6 mm (low-risk group). In contrast, colonoscopy should be performed at 1 year for patients with any large polyp (≥6 mm) or intramucosal cancer (high-risk group). This discrepancy between results of the National Polyp Study may be at least partially explained by the difference in Japanese-style colonoscopy from Western countries, including better bowel preparation using PEG solution taken in the morning on the same day as colonoscopy and meticulous examination using high-resolution colonoscopy and chromoendoscopy in the Japanese-style colonoscopy. Timing of colon preparation might influence the condition of the colon and therefore affect the quality and diagnostic yield of colonoscopy. Colon preparation on the day of colonoscopy achieves a better quality of cleansing and higher diagnostic yield . The quality of colonic cleansing and the detection of flat lesions are significantly improved when the preparation is taken on the day of the colonoscopy . Chromoendoscopy improves adenoma detection compared to white-light colonoscopy, particularly for small and flat neoplasms [29, 49]. On the basis of this perspective, a prospective evaluation to validate international guidelines in the Japanese context seems logical. The Japan Polyp Study (JPS) is a multicenter randomized controlled trial designed to evaluate CRC surveillance strategies in patients who have undergone complete colonoscopies on two occasions, with the removal of all detected neoplasias by high-resolution chromoendoscopy, including the removal of flat and depressed lesions . The study design is similar to the National Polyp Study, but the concept is considerably different using Japanese-style chromoendoscopy. The JPS is intended to continue until 2011 and will help to develop recommendations for surveillance guideline for Japanese patients after polypectomy.
It has become increasingly clear that post-polypectomy surveillance is now a large part of endoscopic practice; however, endoscopic resources and the number of specialists are limited. We also need to consider issues of cost-effectiveness and risk of complications. Based on the available evidence, surveillance guidelines from the United States project that patients with an apparent low risk can wait 5 years, and possibly as long as 10 years, for a repeat colonoscopy. However, further evaluation of this low-risk group is required to confirm the safety of these intervals.
Risk stratification and recommended follow-up intervals are based on the presumption that a high-quality colonoscopy was performed at baseline. However, variable colonoscopic miss rates for adenomas and cancer have been shown [20, 22–25, 27, 28, 40, 42, 70] and surveillance examinations at longer intervals might increase the risk of missing a neoplasm, particularly flat and depressed lesions that may develop into invasive lesions. Consequently, concerns regarding missed neoplasms can be addressed by repeated colonoscopy at a shorter interval. In fact, some reports have shown that a large proportion of endoscopists are conducting surveillance examinations at shorter intervals than recommended in the guidelines . When asked about various polyp scenarios, primary care physicians have also frequently recommended colonoscopy at shorter intervals between examinations than those stated in guidelines .
In Japan, colonoscopic methods, medical economy and resources, and consensus between physicians and patients differ from those in Western countries, so further investigation is needed to establish an appropriate colonoscopy surveillance program for use after polypectomy based on guidelines from the United States.
Conflict of interest
The authors declare that they have no conflict of interest.