Background: Colorectal cancer is the third most common cancer in Sweden, and incidence is increasing. We analysed colorectal cancer detection by colonoscopy in a defined population in Sweden.
Methods: All colonoscopy records for the period 1979-95 in one Swedish county (population 258,000) were retrieved. Information was obtained about patient demographics, date of examination, endoscopists, indications, findings, colonoscopy type and completion level. Records were linked to the Swedish Cancer Register and the Cause of Death Register.
Results: The majority of 2214 colorectal cancers were detected by means other than colonoscopy. In total, 192 were diagnosed and 6 were not detected by colonoscopy, with no significant differences in gender, age, indications, presence of polyps or diverticulosis, time-period or experience of the endoscopist. The mean completion rate of the endoscopist was lower in patients with undetected cancers. Coexisting inflammatory bowel disease was more common in patients with late diagnosis. Sensitivity was 97.0%, and higher when the indication was bleeding, cancer or unclear X-ray findings.
Conclusions: Sensitivity in detecting colorectal cancer was high, and the proportion detected by colonoscopy increased over time. Mean completion rate was lower in patients with undetected cancers. Coexisting IBD was more common in patients with a late diagnosis.