Sigmoidoscopy misses more colorectal cancers

Among older patients, the rate of new or missed left-sided colorectal cancers quadrupled with flexible sigmoidoscopy compared with colonoscopy in a retrospective analysis of 25,541 cases.

The overall interval colorectal cancer (CRC) rate, defined as colorectal cancer diagnosed six-36 months after a lower endoscopy, was 11.7 percent after flexible sigmoidoscopy vs. 2.6 percent after colonoscopy.

The rate of interval CRCs was higher after sigmoidoscopy than colonoscopy in all locations: the descending colon (18.7 percent vs. 3.3 percent), rectum (12.5 percent vs. 2.7 percent), left colon combined (11.7 percent vs. 2.6 percent), sigmoid colon (11 .3 percent vs. 2.4 percent), and rectosigmoid junction (8 percent vs. 2.2 percent).

“Despite the imperfections of colonoscopy, it remains the gold standard for colorectal cancer detection and prevention,” concluded, a fellow with the Mayo Clinic in Jacksonville, Fla.

Although some news reports hailed the study as a boon for colonoscopy, it is unlikely to resolve the long-standing debate over which screening method is optimal. Sigmoidoscopy requires no sedation and less time and bowel preparation for the patient, but cannot visualize the entire colon. Colonoscopy typically requires sedation and has a higher perforation rate, but allows for examination of the entire colon and removal of any detected polyps. Still, these benefits have not translated into an unequivocal reduction in the incidence and mortality from cancer beyond the reach of sigmoidoscopy.

The current study included 25,541 patients 67 years and older at the time of a lower endoscopy during 1988-2005, who were subsequently diagnosed within 36 months with CRC distal to the splenic flexure. All but 841 of the 25,541 cases were detected within the first 6 months.

The patients were identified in the Surveillance, Epidemiology, and End Results-Medicare linked database. Exclusion criteria included participation in an HMO, no Medicare Part B coverage for the 24 months preceding the exam, inflammatory bowel disease, and a history of polyps or family history of colorectal cancer.

Compared with the colonoscopy group, the flexible sigmoidoscopy group was slightly older (78 vs. 77 years), included more women (53 percent vs. 50 percent) and fewer nonwhites (14 percent vs. 16 percent), lived in ZIP codes with higher income/education (no data given), and were more likely to be seen by nongastroenterologists (66 percent vs. 33 percent).

In multivariate logistic analysis, women were at 15 percent higher risk of interval CRC (odds ratio 1.15), while undergoing an inpatient procedure reduced the risk by 47 percent (OR 0.53).

The odds ratio for an interval CRC with flexible sigmoidoscopy was 4.0 (95 percent confidence interval 3.51-4.55).

During a discussion of the study, attendees asked whether detection rates were different among gastroenterologists and nongastroenterologists. Replied that there was no difference (OR 1.09 for nongastroenterologists), but that a difference was observed in the right-side colon in another study presented at the meeting.

That the retrospective study had several limitations including the inability to determine the indication for, or findings of, the lower endoscopy. It also does not apply to patients undergoing screening colonoscopy for detection and removal of polyps, and does not reflect recent advances in endoscopy such as high-definition colonoscopy.