Roel Bogie

Chapter 9

Discussion We observed a more than 50% decrease in PCCRC rate from 2.0 per 1000 colonoscopies before training to 0.8 per 1000 colonoscopies after systematic endoscopist training. The majority of PCCRCs (82%) appeared to be potentially preventable, as they were attributable to missed lesions, nonadherence to surveillance intervals and inadequate bowel examination. Nayor et al. have previously shown that non-adherence to surveillance intervals may increase the number of PCCRC cases, especially in cases where bowel preparation at index examination was inadequate. 17 Apart from non-adherence to surveillance intervals and inadequate bowel preparation, several studies have shown that over 50% of PCCRCs can be attributed to missed lesions and up to 20% to incomplete resection. 6, 7 In line with our results, other studies on education and training in quality of polyp detection and endoscopic resection have shown that colonoscopy quality parameters generally improve after training. 18, 19 Coe et al. 18 have shown that for adenoma detection rate (ADR) and Kaminski et al. 19 for ADR and also for detection of non-polypoid lesions. On the other hand, Shaukat et al. 20 did not observe a significant change in adenoma detection rate of individual endoscopists despite implementation of a training program. Based on previous research, PCCRCs are more likely to be proximally located, smaller in size and have a flat macroscopic appearance compared to prevalent CRCs. Together with the knowledge that adequate bowel preparation is more difficult to achieve in the proximal colonic area, these findings suggest that PCCRCs could have originated from precursors which have been overlooked more easily during index colonoscopy. 6, 7 In the current study, the 11 PCCRC cases we analyzed were equally distributed over the colon. This could be the result of fewer missed proximally located lesions after training. In line with previous data, patients with PCCRC are usually older compared to prevalent CRC cases and have substantially more comorbidity, in particular diverticular disease, 21, 22 whichmakes the performance of a complete colonoscopy far more difficult. Missed lesions were the most common cause of PCCRC (45%) in our study, a finding that is in line with previous data. 3, 7, 23 In another Dutch population-based study on PCCRCs (total n=147) performed by our group, 13.6% of the PCCRCs were attributable to newly developed cancers, 6 comparable with the results in the present study (18.2%). Another preventable factor is inadequate bowel examination (e.g., no cecal intubation or insufficient bowel preparation). Nowadays quality indicators for colonoscopy have been introduced and implemented to assure the quality of each colonoscopic intervention. 4 Incomplete resected polyps have been reported as contributor to PCCRCs in up to 19% of cases. 7, 24, 25 In the present study, no PCCRCs could be attributed to incomplete resection. Non adherence to surveillance interval accounted for 18% of our PCCRCs, a percentage that is higher compared to other studies. 26 It should be noted that adherence to surveillance intervals is a shared responsibility of patients and of physicians and endoscopy staff. 17 Some strengths and limitations of our study should be mentioned. First, this study is unique because of availability of colonoscopy data and CRC data prior to and after systematic training of endoscopists. Second, a detailed description of all PCCRCs is available. In a study on PCCRC, CRC and number of colonoscopies covering the whole South-Limburg region (including three large hospitals), we found that PCCRC rate in the decade prior to systematic endoscopist training

180