Syed G. Shah

Total colonic dye-spray increases the detection of diminutive adenomas during routine colonoscopy: A randomized controlled trial

BACKGROUND Small adenomas may be missed during colonoscopy, but chromoscopy has been reported to enhance detection. The aim of this randomized-controlled trial was to determine the effect of total colonic dye spray on adenoma detection during routine colonoscopy. METHODS Consecutive outpatients undergoing routine colonoscopy were randomized to a dye-spray group (0.1% indigo carmine used to coat the entire colon during withdrawal from the cecum) or control group (no dye). RESULTS Two hundred fifty-nine patients were randomized, 124 to the dye-spray and 135 to the control group; demographics, indication for colonoscopy, and quality of the preparation were similar between the groups. Extubation from the cecum took a median of 9:05 minutes (range: 2:48-24:44 min) in the dye-spray group versus 4:52 minutes (range: 1:42-15:21 min) in the control group (p < 0.0001). The proportion of patients with at least 1 adenoma and the total number of adenomas were not different between groups. However, in the dye-spray group significantly more diminutive adenomas (<5 mm) were detected proximal to the sigmoid colon (p = 0.026) and more patients were identified with 3 or more adenomas (p = 0.002). More non-neoplastic polyps were detected throughout the colon in the dye-spray group (p = 0.003). There were no complications. CONCLUSIONS Dye-spray increases the detection of small adenomas in the proximal colon and patients with multiple adenomas, but long-term outcomes should be studied to determine the clinical value of these findings.

Effect of magnetic endoscope imaging on colonoscopy performance: a randomised controlled trial.

BACKGROUND Colonoscopy can be technically challenging because of unpredictable colonoscope looping. Without imaging, straightening the colonoscope is sometimes difficult since the endoscopist has to guess where the tip is. Magnetic endoscope imaging (MEI), a new non-radiographical technique for picturing the colonoscope shaft in real time, could facilitate loop straightening and thus improve performance. METHODS We assessed trainees and endoscopists with much experience of routine outpatient colonoscopy. In group 1, trainees examined 113 consecutive patients. MEI views were recorded in all examinations, but procedures were randomised to be done by two trainees, either with the endoscopist and endoscopy assistants viewing the imager display (n=58), or without the imager view (n=55). In group 2, two skilled endoscopists were randomised (as with group 1) to undertake consecutive examinations (n=183) either with (n=92) or without (n=91) the MEI view. MEI views of all procedures were analysed retrospectively. FINDINGS In both groups, intubation times were shorter (median 11.8 min [4.3-31.5] vs 15.3 min [4-67] [group 1]; 8.0 min [2.6-40.8] vs 9.3 min [2.5-52.6] [group 2]) and number of attempts at straightening the colonoscope fewer (median 5 [0-20] vs 12 [0-57] [group 1]; 7 [0-55] vs 10 [0-80] [group 2]), when the endoscopist was able to see the imager view. In group 1, colonoscopy completion rates were also higher (100% [58] vs 89% [49]) and duration of looping was reduced (median 3 min [0-18.8] vs 5.4 min [0-44.5]) when the imager could be seen. Abdominal hand pressure was more effective when the endoscopist and endoscopy assistant could see the imager view. INTERPRETATION MEI significantly improves performance of colonoscopy, particularly when used by trainees, or by experts in technically difficult cases; loops were straightened or controlled effectively, resulting in quick intubation times and high completion rates.

Use of video and magnetic endoscope imaging for rating competence at colonoscopy: Validation of a measurement tool

BACKGROUND Counting the number of procedures performed provides at best a crude measure of technical competence in colonoscopy. The aim of this study was to develop and validate a qualitative and a quantitative score for measuring technical competency in colonoscopy using videotape evaluation. METHODS Eighteen endoscopists with varying levels of experience were prospectively videotaped during colonoscope insertion. The following were recorded simultaneously: a closed circuit television view showing instrument handling, the endoscopic luminal view, and a continuous display of the colonoscope configuration (magnetic endoscope imaging). Videotapes were reviewed blindly and in random order by 3 experts. Performance in 3 categories was evaluated: (1) manipulation of instrument controls (0-10), (2) manipulation of the insertion tube (0-6) and (3) depth of insertion (0-4). A global assessment of competence was given for each endoscopist. RESULTS Comparing the total scores as assessed by the 3 blinded experts, for each individual endoscopist, there were significant differences. However, there was good interobserver agreement and correlation between the individual scores and global assessment ratings of competence (p < 0.0001). CONCLUSIONS The video assessment tool described appears to measure technical competence at colonoscopy, although in its present form it lacks reliability. Refinement of the score may improve reliability and deserves further evaluation.

The variable stiffness colonoscope: assessment of efficacy by magnetic endoscope imaging.

BACKGROUND Variable-stiffness colonoscopes combine the flexibility of pediatric instruments for negotiation of the sigmoid colon with the ability to stiffen the insertion tube to prevent or control looping after straightening. Previous studies have found wide variation in the efficacy of the stiffening mechanism. Thus, two studies were conducted to assess the potential benefit of the stiffening device and its optimal use. METHODS In study 1, the effect of routinely stiffening the straightened variable-stiffness colonoscopes in the mid-descending colon was determined in 82 patients. Two insertions were performed (mid-descending colon to cecum) in each patient with and without application of the stiffening device (randomized). The time to negotiate the proximal colon (mid-descending to cecum), time to pass the variable-stiffness colonoscopes across the splenic flexure into the transverse colon, time to pass the right colon, and ancillary maneuvers used were recorded for each insertion. In study 2, consecutive patients, excluding any with previous colonic resection, were examined by using standard adult variable-stiffness colonoscopes. Real-time views of the procedure with magnetic endoscope imaging were recorded for all examinations, but procedures were randomized to be done either with (n = 88), or without (n = 87) the endoscopist viewing the magnetic endoscope imaging display. Whenever stiffening was applied, the anatomic location of the colonoscope tip and stiffness efficacy were recorded. RESULTS In study 1, time taken to negotiate the proximal colon (p = 0.0041) and time to negotiate the splenic flexure (p = 0.006) were significantly shorter and ancillary maneuvers performed were fewer (p = 0.0014) with the stiffening device activated. In study 2, stiffening was used with similar frequency in patients examined with and without the magnetic endoscope imaging view, most commonly for passing the splenic flexure (71%), but also in the transverse colon (12%), right colon (9%), and sigmoid/descending colon (8%). Stiffening was significantly more effective when used in combination with magnetic endoscope imaging (69% with imager vs. 45% without imager; p = 0.0102). CONCLUSIONS Overall, the variable-stiffness device used was effective in controlling looping 57% of the time. Activating maximum stiffness appears to be effective once the sigmoid colon has been negotiated and the colonoscope straightened with the tip in the proximal colon, reducing the number of ancillary maneuvers and shortening the insertion time through the proximal colon. Routine magnetic endoscope imaging further enhances the efficacy of the variable-stiffness colonoscopes by helping to identify the optimal time for stiffening.

4490 Small flat colorectal cancer: experience in 870 consecutive colonoscopies.

INTRODUCTION & METHOD Flat type colorectal cancers have been comprehensively described in the Japanese literature but are thought to occur only rarely in Western countries. We reviewed one endoscopists experience, at a UK centre, in detecting these lesions over a defined period. SUBJECT & RESULT Between Feb.1997 and Oct.1998, one experienced colonoscopist (BPS) performed 870colonoscopies at St Marks Hospital. 45 cases of colonic cancer were found, of which 5 cases (3 male, 2female, mean age 69yrs) were early, flat cancers(11%). Only 1 patient had symptoms (bleeding +pain) directly attributable to the early cancer. Two patients were undergoing surveillance after cancer resection and 1 examination because of previous adenomas. Cancers were located as follows, sigmoid colon 2, transverse colon 1, ascending colon 1, hepatic flexure 1. Size was between 8-15 mm (mean 11mm). Dye-spray was used to define each lesion. Macroscopic appearances, by Kudos classification, three cases were IIc+IIa, one IIa+IIc and one IIa. In no cases was endoscopic resection attempted.All cases had open curative resection, proved histologically to be invasive cancers with no adenomatous component(Dukes A and Jass classification 1). Molecular analysis including APC, P53 and Kras mutation are undertaken. CONCLUSION Small, flat type colonic cancer is a more frequent finding in Western patients than has been previously reported. The lesion found in this study support the possibility of de-novo caricinogenesis occuring in some patients.

3460 Endoscopic dilatation of crohn's strictures: longterm outcomes in 85 consecutive patients.

Background: Endoscopic dilatation of Crohns strictures is a minimally invasive technique that may avoid surgery and preserve intestinal length. Longterm outcomes after dilatation require clarification. Aims: To review all dilatations of Crohns strictures performed at a centre with a specialist interest in colonoscopy. Methods: Case notes of all Crohns patients who underwent endoscopic stricture dilatation between 1988-1999 were reviewed. Results: 89 patients were identified, with follow-up information available in 85 patients (45 female,166 dilatations). Median age at first dilatation was 42(23-74). 109(65%) were performed as day cases with conscious sedation and 4 (2%) under GA. Median dilatations per patient was 1 (1-8). Strictures were anastomotic (ileo-colonic 50, ileal 6, ileo-rectal 4, sigmoid 2, pouch-ileal 1) in 63 patients (133 dilatations) and denovo (ileal 7, ileo-caecal 2, transverse 1,left colon 9, rectum 3,) in 22 (33 dilatations), median stricture length 2.5cm (1-12.5). Maximum balloon size was recorded in 125, median 18mm (10-20). In 10 distal strictures a rigid dilator (median 16mm [12-18]) was used after endoscopic guidewire placement. Median follow up time was 29.1 months (0.23-190.2). 67 dilatations (40%) were followed by definite clinical improvement. 21 patients (25%) have had no further procedure since their first dilatation, median follow up 23.6 months (0.3-102). 35 (41%) patients had a further procedure within 6 months of first dilatation (surgery 19, dilatation 16). Of these, 13 were technical failures at dilatation. Multiple dilatations were performed in 31 patients (36%), median interval 21.3 weeks (0.1-537.6). 44 patients (52%) had surgery for the stricture during follow up. 7 complications were identified (4.1%). 4 patients (2.9%) with dilatations for anastomotic strictures had minor bleeding not requiring intervention. 3 patients (8.6%) with denovo strictures had complications: 2 left colonic perforations (1 temporary loop ileostomy formed, 1 managed conservatively); the other (ileal) had a minor bleed. There was no mortality related to dilatation. Conclusions: Endoscopic management achieves long-term benefit in approximately 50% of patients with accessible Crohn s strictures. Repeat dilatation may be necessary. Most dilatations can be performed in endoscopy with conscious sedation. Greater caution is required when dilating de novo strictures.

Performance in Colonoscopy Can Be Reliably Scored in the First 8 Minutes of Insertion: Further Evidence for a Tri-split Video Assessment Tool

Performance in Colonoscopy Can Be Reliably Scored in the First 8 Minutes of Insertion: Further Evidence for a Tri-split Video Assessment Tool Siwan Thomas-Gibson, Syed Shah, David Swain, Catherine Thapar, Gillian Schofield, Brian Saunders Background: We have previously described use of a tri-split video assessment tool to score colonoscopy performance and highlight specific flaws in technique. The aims were 1) to investigate if scoring just the first 8 minutes of insertion was adequate and 2) to test inter-rater agreement (reliability) of the revised performance score. Methods: Ten doctors training in colonoscopy were videoed. The trainer rated all cases as ‘moderately easy’ at the time of the procedure. ’Trisplit’ views (simultaneous views of the endoscopists’ hands, mucosal view and endoscopic imager view) from each case were recorded. Two cases from each endoscopist (total 20)were edited to include only the first 8mins of insertion. These were then edited together in random order. A second tape of the entire insertions from the same cases (different randomisation) was also made. Two experienced endoscopists blinded to the doctors’ identity, independently scored the procedures using amodified version of a score sheet previously described. A Cumulative score (range 5-15) for 5 parameters and an overall Global score (range 1-3: Not competent, Reasonably competent, Fully competent) were given for each case based on the tri-split views. The 5 parameters were: Correct grip of instrument head; Tip steering/maintenance of mucosal view; Appropriate use of suction/ insufflation; Correct grip/manipulation of instrument shaft; Anatomic depth of insertion. Results: There was good correlation (*Pearson’s correlation coefficient) between scores for the 8min and entire insertions suggesting 8mins is long enough to score insertion technique. Scorer 1: Cumulative scores r* = 0.644 p=0.002; Global scores r* = 0.7369 p= 0.0002. Scorer 2: there was a strong correlation between the Global scores for the 8 min and entire insertion tapes: r*=0.6577 p=0.0056 (Cumulative scores not significant). There was moderate inter-rater agreement for Global scores in the 8 min insertions k= 0.46. Conclusions: This provides further evidence for a feasible method of remote objective performance assessment in colonoscopy using tri-split video to highlight specific flaws in technique. It is valid to score performance using only the first 8 minutes of insertion. An overall Global assessment of competence derived from the parameters scored using a tri-split view is most reliable.

3363 Video evaluation of endoscopic competency.

Background: Previous attempts at measuring competency in colonoscopy have scored depth of endoscope insertion, procedure time, identification of anatomical landmarks & frequency of ancillary manoeuvres used. Instrument manipulation/dexterity & hand-eye co-ordination are key skills necessary for achieving proficiency, but have never been objectively evaluated. Aims: To develop a qualitative & quantitative score for measuring technical competency in colonoscopy using videotape evaluation. Methods: Eight endoscopists at varying levels of experience were prospectively videotaped performing colonoscopic insertion & withdrawal, using a purpose designed video tri-split mixing unit. Three images were generated, 1). a CCTV view showing instrument handling, 2). the endoscopic mucosal view, and 3). a continuous display of anatomical depth of insertion using magnetic endoscope imaging (MEI). The video tapes were reviewed blindly and in random order by 2 experts who independently scored 4 aspects of the examination: 1. Manipulation of instrument controls [instrument grip (0-4), tip/steering control (0-3) & appropriate aspiration/insufflation (0-2)] 2. Manipulation of instrument shaft [Grip (0-2), rotation/torqueing (0-6), attempts to straighten scope (0-6)] 3. Mucosal visualisation (0-6) 4. Anatomical depth of insertion, within first 5 mins of the examination (0-4)(Maximum total score summed=33). Results: Comparing the ratings/scores as assessed by the 2 blinded experts, for each individual endoscopist, there was good repeatability/agreement for the total score summed for each of the 4 criteria (mean difference 1 (SD 3.2); 95% limits of agreement [-5.4 to 7.4]) (See Table). Conclusions: Video-evaluation provides both a reliable & objective measurement of endoscopic competence, and provides specific feedback to trainees. Further work is necessary to evaluate its usefulness in monitoring training/performance & developing standards of practice.

7215 Individualising a patient's need for sedation prior to colonoscopy.

Background: In the UK most patients having colonoscopy are pre-medicated with intravenous sedation to relieve anxiety & alleviate pain. Recent studies suggest that selective sedation is feasible and acceptable to patients. Aims: To examine those factors which are thought to influence analgesia requirements during colonoscopy, with a view to developing a rank score to predetermine sedation needs prior to colonoscopy. Methods: Consecutive out-patients undergoing routine colonoscopy were invited to participate. All patients completed a questionnaire, from which the following demographic and clinical features were recorded: age, gender, occupation, ethnicity, current use of anxiolytics/analgesics, obstetric history/parity, previous pelvic/abdominal surgery, and previous colonoscopy experience. Anxiety was assessed using the seven item HAD sub-scale and a 100mm VAS. Severity of discomfort associated with bowel prep was also scored (100mm VAS). Prior to colonoscopy all patients were given a combination of low-dose midazolam (1.25mgs) & pethidine (25mgs). Post-procedure patients scored level of pain/comfort/satisfaction (100mm VAS) & willingness to repeat colonoscopy with/without sedation. Results: 120 patients were studied, 47 (39%) showed a willingness to try without sedation for future colonoscopy, 66 (55%) expressed a preference for sedation, and 7 (6%) were undecided. Univariate logistic regression analysis indicated that increasing age, male gender, low anxiety, and clinical indication were predictive of willingness to try colonoscopy without sedation in the future. However, when these factors were incorporated into a multiple logistic regression model, only increasing age (odds ratio for 1-year increase 1.03 [1.007-1.062]; p=0.012), male gender (odds ratio 2.92 [1.29- 6.62]; p=0.010) and anxiety (HAD score) (odds ratio for a one unit increase in HAD score 0.88 [0.80-0.97]; p=0.017) were predictive of non-sedation. Patient pain scores were lower, and procedure tolerability/satisfaction greater in the group willing to try without. Conclusions: In this study, increasing age, male sex & low anxiety correlate with a willingness to undergo unsedated colonoscopy in the future.We plan to develop and validate a simple scoring system, which will allow us to recommend appropriate sedation/non-sedation based on these findings.

⁎⁎3640 Magnetic endoscope imaging & colonoscopy performance in trainees.

Background: For trainees, colonoscopy remains one of the most difficult endoscopic techniques to perform. This is largely due to the huge variability in colonic anatomy, which makes accurate appreciation of instrument looping & tip localisation difficult to judge. Aims: To evaluate the effect of magnetic endoscope imaging (MEI) on performance of colonoscopy by trainee endoscopists. Methods: Consecutive out-patients undergoing colonoscopy were studied. Patients with previous colonic resections were excluded. Two experienced trainees were randomised to perform consecutive examinations either with or without the MEI view. The efficacy of abdominal pressure was accurately assessed using a single sensor coil attached to the endoscopy assistants hand. The magnetic imager view of each procedure was recorded on computer disk and retrospectively analysed. The clinical findings, total intubation time & extent of intubation, and patient pain score (100mm visual-analogue scale (VAS), 0=none, 100=very severe) were also recorded. Results: 94 patients were studied, 48 with and 46 without the imager view. Total colonoscopy was achieved in all cases except three (blinded to imager view). Frequency of looping was similar in the study groups, although atypical loops were more common in patients examined without the imager view (14/46 vs 1/48; p=0.0001 by Fishers exact test). The duration of loop formation per patient, however, was significantly less when the trainee was able to see the imager view (median 2.6 mins (0.2-18.7) vs. median 5.0 mins (0.2-27.1); p=0.0056 by Mann-Whitney test) as was the number of attempts taken to straighten the colonoscope per procedure (median 5 (0-19) vs. median 10 (0-57); p=0.001 by Mann-Whitney test). Comparing intubation time in the two study groups, the procedure was also significantly quicker with the benefit of the imager view (median 10.4 mins (4.3-27.4) vs. median 13.3 mins (4-50); p=0.013 by Mann-Whitney test). Abdominal pressure was also more effective when the endoscopist and endoscopy assistant could see the imager view (beneficial 45/62 vs. 19/45; p=0.0026 by Fishers exact test). Patient pain scores were similar in the two study groups. Conclusions: In trainees, magnetic endoscope imaging significantly improves performance of colonoscopy, making the procedure quicker & allows loops to be accurately assessed and straightened more effectively.