Endoscopic perforation of the colon: lessons from a 10-year study.

OBJECTIVES: To assess the incidence, clinical features, and management of endoscopic colon perforations in a large number of patients at a major medical teaching center.
METHODS: A retrospective review of medical records of all patients with colon perforations from endoscopy over a 10-yr period.
RESULTS: A total of 10,486 colonoscopies were performed over a 10-yr period. There were 20 (0.19%) perforations and two (0.019%) deaths related to colonoscopy and two perforations with no deaths in 49,501 sigmoidoscopies (0.004%). The majority of perforations (65%) occurred in the sigmoid colon. The mean age of the patients was 72 yr (range, 48-87 yr). Multivariate analysis using gender and age showed that female gender was an independent predictor of a higher risk of perforation (p < 0.05). Electrocautery injury (36%) and mechanical injury (32%) from the tip and shaft of the endoscope were the major causes for perforation. Most patients (91%) presented within 48 h of endoscopy. Nine patients (47%) required a surgical resection with primary anastomosis; seven (37%) required a simple closure. The average hospital length of stay was 7.7 +/- 2.8 days. Although trainee endoscopists were involved in only 20% of the colonoscopies performed, eight (40%) perforations occurred while the training fellow was involved in the case. However, this increased risk of perforation with a training fellow was not statistically significant (p = 0.625).
CONCLUSIONS: Colonoscopy can result in significant morbidity and carries a small risk of death. Sigmoidoscopy has lower risk. The following situations may represent increased risk to colonoscopy patients: unusual difficulty in traversing the sigmoid colon; difficult examinations in female patients, and difficult examinations performed by trainee physicians.