Predicting proximal advanced neoplasms at screening sigmoidoscopy.

Abstract

This study was designed to assess the predictive value for advanced proximal neoplasms (cancer, adenoma > or = 10 mm, or villous component > 20 percent, or severe dysplasia) of the characteristics of distal polyps. The study was conducted among patients, aged 55 to 64 years, referred for colonoscopy in the Italian trial of sigmoidoscopy screening for colorectal cancer. Patients reporting a history of colorectal cancer, adenomas, inflammatory bowel disease, recent colorectal endoscopy, or two first-degree relatives with colorectal cancer were excluded. We compared the prevalence of advanced proximal neoplasia in patients with "low-risk" (1-2 tubular adenomas, < 10 mm, with low-grade dysplasia, or hyperplastic polyp) and in those with "high-risk" (size, > or = 10 mm, or > or = 3 adenomas, or villous component > 20 percent, or severe dysplasia) polyps in the distal colon. Of 426 patients with polyps > 5 mm, 29 (6.9 percent) were detected with an advanced proximal neoplasm (including 4 colorectal cancers). The prevalence of proximal advanced neoplasia was 9.4 percent among patients with high-risk distal polyps and 2.5 percent among those with low-risk lesions (adjusted odds ratio, 3.19; 95 percent confidence interval, 1.06-9.59). Approximately 40 people with low-risk distal polyps 6 to 9 mm should undergo colonoscopy to detect one proximal advanced neoplasm; the corresponding number for patients with high-risk distal polyps is 10. The 2.5 percent prevalence of proximal advanced neoplasms among people with low-risk 6-mm to 9-mm distal polyps is similar to the prevalence observed among people without distal polyps. Restricting colonoscopy referral to patients with high-risk distal polyps might represent a cost-effective strategy in a screening context.