Video Capsule Endoscopy for the Evaluation of Small Bowel Polyps in Familial Adenomatous Polyposis (FAP)

Abstract

Purpose: Compare video capsule endoscopy (VCE) to standard endoscopy for the diagnosis of small bowel polyps. Methods: This was a single center, prospective study. Eligible patients had a clinical diagnosis of FAP. Most patients were concurrently enrolled in a NIH study and had at least 20 polyps in the duodenum and prior placement of a tattoo in the proximal small bowel (n = 28). Four FAP patients referred for routine VCE evaluation for small bowel polyps were also included. VCE was performed followed by push enteroscopy (PE) and lower endoscopy (LE). Two VCE readers independently determined the number and sizes of polyps in the following regions of small bowel: 1) defined segment of small bowel from the duodenal bulb to the tattoo and 2) quarters of small bowel determined by small bowel transit time. A single endoscopist blinded to VCE determined the number and sizes of polyps in the following regions: 1) defined segment of small bowel from the duodenal bulb to the tattoo; 2) maximum insertion length of PE; and 3) maximum insertion length of LE. The primary outcome was the number of polyps in the defined segment of small bowel, comparing VCE to the gold standard of PE. Agreement was calculated utilizing kappa coefficients. Results: Thirty two patients participated in the study (median age 45 years, 41% female). On VCE, 94% of patients had polyps proximal to the tattoo compared to 100% on PE. VCE diagnosed less polyps in this region (median 10 and 9 for each reader) compared to PE (median 41) (p = 0.002). Agreement between VCE and PE was fair for polyp counts (k = 0.34, 0.36). Agreement for determining the size of the largest polyp was poor to fair (k = 0.10, 0.22). VCE and PE disagreed for detecting large polyps (≥1cm) (k =−0.20, −0.27). For the entire small bowel, VCE detected significantly less polyps (median 38 and 54 for each reader) compared to combined (PE+LE) endoscopy (median 138) (p < 0.001). Agreement was fair to moderate between VCE and combined endoscopy (k = 0.21, 0.56). On VCE, the percentage of patients with polyps in the first, second, third and last quarter of small bowel was: 100%, 35%, 15% and 46%, respectively. Polyps were more common in the first quarter than in the remaining quarters combined (p < 0.0001). Mean size of polyps was less than 5 mm in each quarter. Conclusions: VCE underestimates the number of small bowel polyps in patients with FAP and does not reliably detect large polyps.