High-Quality Colonoscopy: Withdrawal Techniques

For complete visualization of the colon, intubation of the cecum is a necessity, and therefore proper scope and loop handling are basic skills required for a high-quality colonoscopy. Although the criterion cecum intubation rate of greater than 90% is often met in daily clinical practice, it is known that the ADR differs significantly between endoscopists. This difference is partly explained by differences in the withdrawal technique. A video analysis of colonoscopy withdrawals performed by two endoscopists at opposite ends of the spectrum of ADR showed that the endoscopist with the higher ADR visualized a greater percentage of the colonic mucosa because this endoscopist (1) was more adequate in examining the proximal side of the haustral folds, flexures, and rectal valves; (2) was more thorough in removing remnant fluid or fecal material; (3) distended the colon better; and (4) spent more time viewing the mucosa. This was confirmed in another study by Lee and colleagues showing that moderate-level (ADR between 21% and 42%) or high-level (ADR >42%) detectors had higher scores on their withdrawal technique after video analysis of 20 colonoscopies of each of 11 participants, all working in different facilities. The most important differences were found between low-level detectors and moderate- or high-level detectors, suggesting that proper withdrawal techniques are required for higher adenoma detection. Nonetheless, other skills, such as pattern recognition, may define even more the difference between moderate- and high-level adenoma detectors. Moreover, no difference was observed in withdrawal time between low- and high-level detectors. Increased detection of adenomas has, however, been associated with a withdrawal time of greater than or equal to 6 minutes, and this quality indicator is currently endorsed as a surrogate marker of the quality of withdrawal technique. Nonetheless, introducing a protocol with a minimum time for withdrawal greater than or equal to 6 minutes has shown conflicting results with regard to an increase in ADR. It is therefore likely that factors other than withdrawal time alone influence the ADR, and withdrawal skills in all aforementioned domains need to be improved.

The question of whether improving the withdrawal technique results in a higher ADR is still unanswered. Raising awareness of quality measurement by video recording has been shown to improve withdrawal techniques, but had no effect on ADR. The study was, however, underpowered for this end point. Endoscopists randomized to a training program (EQUIP training) increased their ADR from an average of 36% to 47%, compared with control subjects, which remained at an ADR of 35%. The EQUIP program consisted of training in the importance of good practice withdrawal techniques, but also of training in pattern recognition. This study showed that endoscopists are able to improve their ADR, but the extent to which it can be increased by only improving the level of performance in all domains of the withdrawal technique remains unknown. It seems likely that improving withdrawal techniques improves visualization of the colonic mucosa, and therefore will increase the ADR at least to some extent.

Retroflexion in the Right Colon

Retroflexion of the endoscope in the rectum to visualize the distal rectum and dentate line is a widely accepted and performed standard part of routine colonoscopy. Retroflexion can also be performed in the proximal colon. Because of the diameter of the right colon, retroflexion can be achieved in most patients, and failure to do so is mainly caused by looping of the insertion tube. It was hypothesized that retroflexion in the proximal colon could increase the detection of adenomas located on the proximal sides of haustral folds that are difficult to detect by forward viewing. A randomized study, comparing a second pass of the proximal colon in forward-viewing position in 48 patients with a second pass in retroflexed position in 50 patients, showed an adenoma miss rate of 33.3% in forward view versus 23.7% in retroflex view. Based on these findings the authors concluded that retroflex view of the proximal colon was of no additional value. This was indirectly supported by a prospective, observational study of 1000 patients that all had a second pass in the right colon in retroflex view. This study showed a polyp miss rate of 9.8%, which was much lower than the miss rate of 27% for proximal colon polyps reported in tandem colonoscopy studies. Although retroflexion in the right colon should be performed carefully to prevent mechanical damage to the bowel wall, no adverse events were reported in either study. Based on the aforementioned studies, routine retroflex view of the proximal colon is not expected to increase adenoma detection compared with a second pass of the proximal colon in forward view. Because the detection of a proximal neoplasia at first pass was an important predictor of missed polyps in the proximal colon, a second pass in forward view may be useful in those patients with a proximal neoplasia detected. This strategy has, however, not been formally tested.

Use of Antispasmodic Agents on Colonic Surface Area Visualization

Besides remnant fecal material, haustral folds, and sharp bends, colonic motility can interfere with good visualization of the colonic mucosa. It has been hypothesized that inhibiting smooth muscle contractions during withdrawal may increase mucosal visualization, and thereby polyp and adenoma detection. Hyoscine- N -butylbromide has a superior safety profile and costs less than glucagon, making it the preferred agent to inhibit colonic motility. Although increased polyp detection was reported with hyoscine- N -butylbromide in retrospective and prospective cohorts, recently published randomized placebo-controlled studies did not show a significant increase in ADR with regular use of hyoscine- N -butylbromide. Lee and colleagues found a nonsignificant increase in polyp detection (1.2 vs 0.41; P = .06) in patients with more than grade 3 spasm of the colon during insertion. The observed difference was small, however, and the ADR was not mentioned. Two meta-analyses on the use of hyoscine- N -butylbromide did not show a significant increase in polyp or adenoma detection between hyoscine- N -butylbromide and placebo. In a recent randomized controlled study of hyoscine- N -butylbromide including 402 participants, visualization of nonpolypoid lesions was higher in the placebo arm of the study (1 vs 11) ; however, this finding was not confirmed in another study (Ter Borg F, personal communication, 2014). Based on these observations, the routine use of hyoscine- N -butylbromide is of no benefit for mucosal visualization, and there is concern whether this hampers the visualization of nonpolypoid lesions.

Positioning of the Patient

Adequate luminal distention is essential for full visualization of the colonic mucosa. This can be partly achieved with insufflation of room air or CO ₂ , but may be unsuccessful as the air travels to proximal segments of the colon leading to excessive insufflation, which is associated with abdominal symptoms during or after colonoscopy. Position changes have been shown to increase visualization at virtual colonoscopy. In addition, in a video analysis of tandem colonoscopies, luminal distention of several segments of the colon was scored during different positions. Better luminal distention was achieved with the left lateral position for the cecum toward the hepatic flexure, the supine position for the transverse colon, and the right lateral position for the splenic flexure and the descending colon ( Fig. 1 ). This prompted the authors of this study to perform a randomized crossover trial to validate the effect of better luminal distention on the ADR in the proximal colon. Patients were randomized to withdrawal of the colonoscope in left lateral position only or in the previously summarized dynamic sequence. Each segment was examined in tandem fashion. Dynamic positioning was associated with an 11% increase in ADR from 23% to 34%, with 61 adenomas detected with dynamic positioning compared with 42 adenomas detected with standard left lateral positioning. The differences in polyp and adenoma detection were particularly observed in the transverse colon, splenic flexure, and descending colon, with the most apparent improvement in the transverse colon. Another study with comparable design reported an increase in ADR from 24% to 33% with dynamic positioning with an increase from 42 to 53 detected adenomas. A more recent study in 776 patients comparing dynamic positioning with standard care did not confirm these results. Several explanations can be provided for these conflicting data. In the study by Ou and colleagues, position changes were allowed in the control arm of the study, and 32% had at least one position change during endoscope withdrawal. Furthermore, ADR was 40% in the dynamic positioning group against 39% in the control group. The authors suggested that the benefit of position changes may be of less value when appropriate withdrawal techniques are applied. Despite these contradictory results, it seems reasonable to assume that position changes can be of added value for the endoscopist as one of the options to improve distention of the colon segments when standard options, such as insufflation, are unsuccessful.

Position changes used during extubation.

( From East JE, Bassett P, Arebi N, et al. Dynamic patient position changes during colonoscope withdrawal increase adenoma detection: a randomized, crossover trial. Gastrointest Endosc 2011;73(3):458; with permission.)

Removing Polyps During Insertion

Prevention of postcolonoscopy interval colorectal cancer is not only related to detection of adenomas, but also their removal. It is recognized that small polyps visualized during insertion may be difficult to find during withdrawal. Several explanations have been postulated, such as a different anatomic position of the colon during insertion of the endoscope, exposing other parts of the mucosal surface. In a prospective study of 368 patients, 61.5% of the nonadvanced adenomas were already detected during insertion with an additional detection of 38.5% nonadvanced adenomas during withdrawal. Most polyps were detected in the sigmoid during insertion. A study randomized patients between polypectomy of less than 10-mm polyps during insertion and withdrawal (N = 150) or during withdrawal only (N = 151). This study concluded that adenomas were missed in 7.3% of patients when removing polyps only during withdrawal of the endoscope. These missed polyps were primarily located in the sigmoid and descending colon. This considerable adenoma miss rate suggests that small polyps should be removed during insertion. However, two recent studies, randomizing patients between polypectomy during insertion and withdrawal versus during withdrawal only, showed no benefit of removing small and diminutive polyps during insertion. There were, however, some major differences between the studies. The study by Hewitt and colleagues evaluated the effect of an additional 3 minutes inspection on ADR during insertion, whereas removal of polyps during insertion was also allowed in the control arm consisting of inspection during withdrawal only. In contrast, in the study by Sanaka and colleagues polyps were only removed in the withdrawal phase of the colonoscopy. In both studies, most colonoscopies were performed by high-level detectors (ADR ranging from 34% to 58%) in populations with a different case mix of colonoscopy indications. Notwithstanding these conflicting data, it seems reasonable to assume that removing all small polyps detected during insertion, especially diminutive polyps in the sigmoid or descending colon, may be beneficial.