Long-term Follow-up Study Designed to Evaluate the Relative Risk of Two Colonoscopy Schedules for Patients With Small Polyps

Primary Hypothesis: Risk factors can be determined for large (>1 cm) adenomas, precursor
lesions for colorectal cancer.

Secondary Hypothesis: Determine long-term rates for development or recurrence of polyps;
determine sensitivity/specificity of current colon cancer screening strategies; determine
relationship of dietary factors and biomarkers of cell proliferation; determine the efficacy
and safety of long-term (5 years) repeat colonoscopy in patients with small polyps.

Intervention: Phase I: All patients undergo full colonoscopy. Phase II: Randomization to
repeat colonoscopy at 2-3 years and 5 years after baseline, or, repeat colonoscopy at 5 years
only. Phase III: Ten-year follow-up on all Phase I patients for medical outcomes. Repeat
colonoscopy at 10 years on polyp-free patients (Phase I) aged 50-64.

Primary Outcomes: Presence of risk factors and adenomatous polyps including prevalence,
descriptive characteristics, and long-term occurrence/recurrence rates.

Study Abstract: Phase I is a cross-sectional study designed to identify risk factors for
large (>1 cm) adenomatous polyps. Approximately 3200 asymptomatic subjects (age 50-75) have
completed risk factor assessment, medical and dietary histories, and have undergone complete
colonoscopy examination. This will identify for comparison purposes a polyp-free control
group and is the first large prospective study to include such a group. Data at colonoscopy
will characterize the prevalence, size and distribution of adenomatous polyps. This will
permit an assessment of sensitivity of sigmoidoscopy in this population. In addition, tissue
from normal rectal mucosa will be analyzed for evidence of cell proliferation activity. The
primary focus of Phase I is a risk factor analysis. A multivariate analysis will be performed
to determine the relationship of historical and environmental factors as well as cell
proliferation activity with the presence of adenomatous polyps. A cohort consisting of a
subgroup of polyp patients (large and small) and matched polyp-free controls will be tracked
longitudinally to determine polyp occurrence/recurrence rates.

Phase II of the study is a long-term follow-up study designed to evaluate the relative risk
of two repeat colonoscopy schedules for patients with small polyps identified in Phase I of
the study. Recruitment is complete with 615 patients eligible (of the target 808) assigned at
random to either repeat colonoscopy at 2-3 years and 5 years, or to repeat colonoscopy at 5
years only. This phase will also provide preliminary longitudinal risk factor information
related to occurrence/recurrence of polyps.

Phase III was a 5-year extension of follow-up period. All Phase I patients were to be
reconsented to provide medical outcome data for a period of 10 years from baseline exam.
Phase I patients polyp-free, aged 50-64 will be offered repeat colonoscopy at 10 years to
evaluate long-term risk.

Results (Phase I): 3121 patients had complete colonoscopy which revealed high rates of
neoplasia: 37.5% had one or more neoplastic lesions; 10.5% had advanced neoplasia including
30 cases of invasive cancer (1%). There were 3.7% of patients with no lesions in the rectum
or sigmoid colon who had advanced neoplasia elsewhere in the colon: 32% of all patients with
advanced neoplasia would not be detected with an exam of the rectum or sigmoid colon
(distal); 62% of patients with proximal advanced neoplasia would not be detected with an exam
of the rectum and sigmoid colon. There were few serious complications (0.3%).

The one-time fecal occult blood test (FOBT) was evaluated as a diagnostic test for advanced
neoplasia. A positive FOBT indicated an increased likelihood (3-4x) of advanced neoplasia.
However, one-time FOBT failed to detect 75% of patients with advanced neoplasia.

The primary analysis of risk factors (Phase I) found positive associations (for advanced
neoplasia) for history of a first degree relative with colorectal cancer (OR, 1.66; 95% CI,
1.16-2.35), current smoking (OR, 1.85; 95% CI, 1.33-2.58), and current moderate to heavy
alcohol use (OR, 1.02, 95% CI, 1.01-1.03). Inverse associations were found for cereal fiber
intake (OR, 0.95, 95% CI, 0.91-0.99), vitamin D intake (OR, 0.94, 95% CI, 0.90-0.99), and use
of NSAIDs (OR, 0.66, 95% CI, 0.48-0.91). Results appeared in JAMA (2003;290:2959-2967).

Phase III is completed with patients completing their scheduled follow-ups. A major
manuscript on the sensitivity/specificity of digital rectal exam appeared in Annals of
Internal Medicine January, 2005. The Phase II results manuscript appeared in Gastroenterology
in October 2007.

Inclusion Criteria:

Study Complete

Exclusion Criteria: