Is routine ileoscopy useful? An observational study of procedure times, diagnostic yield, and learning curve

IIeoscopy is not routinely attempted because of its perceived technical difficulty, time constraints, and the expectation of a low diagnostic yield.

To investigate the value of routine ileoscopy as an integral part of colonoscopy in terms of additional diagnostic information, extra time spent, and the relationship between ileoscopy rate and accumulation of colonoscopic experience.

We examined colonoscopy data from September 1995 to April 2004 of a gastroenterological firm. Crude and adjusted total colonoscopy rates (CTCR, ATCR) and ileoscopy rate (IR) were calculated. For calculation of ATCR and IR, 108 procedures in patients with previous colonic resection and 91 with unavoidable reasons for failure to reach cecum were excluded. Time trend in ileoscopy rate was analyzed with IR as a function of cumulative colonoscopy experience. Data on procedure times were collected for 1,222 consecutive colonoscopies between November 2000 and April 2004. Sixty-three procedures in patients with previous colonic resection, and 47 unavoidable and 14 avoidable failures to reach cecum were excluded from analysis of procedure times.

Of 2,537 colonoscopies, 1,902 were performed by a single consultant and 635 by eight trainees with or without assistance from the consultant. The CTCR, ATCR, and IR were 94%, 97.5%, and 71.5% respectively. IR and ATCR rose progressively to plateau at 85% and 99% after 600 and 750 procedures respectively. The diagnostic yield from ileoscopy and ileal histology was 16.7% and 19% in patients with colonic inflammatory bowel disease, and 2.69% and 7.4% in other patients. Twenty six ileoscopies in 24 patients showed Crohn's ileitis with normal colon. The other diagnoses were NSAID related or nonspecific ileitis, ileal lipoma, ileal villous atrophy, and amyloidosis. The median anus to cecum and cecum to ileum times were 8.5 and 2 minutes (interquartile ranges=5.5 to 14 and 1 to 4 minutes) respectively. Procedure times for colonoscopies involving trainees were significantly longer compared with those by the consultant alone (median anus to cecum and cecum to ileum times=16.5 v 7 and 2.5 v 1.5 minutes; p<0.0001 and 95% CI for difference=7.5 to 9 and 0.5 to 1 minutes respectively). The median length of ileum examined was 15 cms (IR=10-20 cms).

IIeoscopy is the gold standard in the documentation of completeness of colonoscopy. With practice, it can be achieved routinely in at least 85% of colonoscopies. In skilled hands, it adds only 3 minutes to the procedure time, and contributes significantly to quality assurance and diagnostic yield.