A prospective analysis of 13,580 colonoscopies. Reevaluation of credentialing guidelines.

Background: The aims of this study were to assess the safety and efficacy of surgeons performing colonoscopy, and to use the results to reevaluate currently available credentialing guidelines.

Methods: A prospective outcomes study was designed to include all members of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES). End points were related to the efficacy and safety of colonoscopy. Credentialing guidelines were reviewed.

Results: Between April 1998 and September 1999 13,580 colonoscopies were prospectively entered into a database. The most common indications were rectal bleeding, colonic polyps, and change in bowel habits. The colonoscopy was normal or revealed only diverticulosis or nonspecific inflammation in 8,473 (62.4%), lower gastrointestinal bleeding in 4 (0.03%), polyps in 4,645 (34.2%), and tumors in 458 (3.4%) patients. The most common biopsy methods for polyps or tumors were the snare (n = 1,728; 34%), the hot (n = 1,600; 31%), and the cold (n = 1,340; 22%) procedures. The colonoscopy was complete in 12,495 cases (92%), requiring a mean procedure time of 22.7 min (range, 1-170 min). Intraprocedural complications included arrhythmia (n = 14; 0.1%), bradycardia (n = 115; 0.8%), hypotension (n = 171; 1.2%), and hypoxia (n = 806; 5.6%). Postprocedural complications were seen in 27 patients (0.2%). Bleeding (n = 10; 0.07%) was managed by observation alone (n = 9; 0.06%) and repeat colonoscopy with transfusion (n = 1; 0.01%). Perforation (n = 10; 0.07%) was treated successfully by observation with conservative management (n = 5; 0.05%) and surgery (n = 5; 0.05%); severe abdominal pain (n = 4; 0.03%) was managed by observation and conservative therapy; and bronchospasm (n = 2; 0.015%) was managed by observation and supportive care. One single mortality (0.007%) was that of a 70-year-old man with a massive lower gastrointestinal hemorrhage who had a cardiac arrest in the recovery room following colonoscopy. The complication rate was not significantly associated statistically with either the level of experience or the number of prior or annual colonoscopies. However, prior colonoscopic experience did have an impact on the completion rate (p < 0.001) and was inversely proportional to the time to completion (p < 0.001). Similarly, the number of annual colonoscopies affected the completion rate and was inversely correlated with the time to completion (p < 0.001).

Conclusions: This large prospective outcomes study showed that colonoscopy performed by surgeons can be rapidly and successfully done with acceptably low morbidity and mortality. There was no association between experience and complications. However, a minimum of 50 prior colonoscopies and 100 annual colonoscopies were associated with a significant improvement in the rate of completion. There was also a significant correlation between both prior and ongoing annual experience and the time required for the examination. No minimum number of cases can be mandated for credentialing to perform "safe" colonoscopies.