Background: Guaiac-based fecal occult blood tests (FOBTs) for colorectal cancer screening are not specific for human hemoglobin and have low sensitivity. Automated-development, immunochemical FOBT is quality-controlled, is specific for human hemoglobin, and does not require diet restriction.
Objectives: To measure the sensitivity and specificity of quantitative immunochemical fecal hemoglobin measurements for detection of cancer and advanced adenoma in patients undergoing colonoscopy, to determine fecal hemoglobin thresholds that give the highest posttest probability for neoplasia, and to determine the number of immunochemical FOBTs needed.
Design: Prospective, cross-sectional study.
Setting: Ambulatory endoscopy services of the main health medical organization in Tel Aviv, Israel.
Participants: 1000 consecutive ambulatory patients--some asymptomatic but at increased risk for colorectal neoplasia and some symptomatic--who were undergoing elective colonoscopy and volunteered to prepare immunochemical FOBTs.
Intervention: The hemoglobin content of 3 bowel movements was measured, and the highest value was compared with colonoscopy findings.
Measurements: Sensitivity, specificity, predictive values, likelihood ratios, and 95% CIs of fecal hemoglobin measurements for clinically significant neoplasia, their relationship to the amount of fecal hemoglobin measured, and the number of immunochemical FOBTs performed.
Results: Colonoscopy identified clinically significant neoplasia in 91 patients (cancer in 17 patients and advanced adenomas in 74 patients). Using 3 immunochemical FOBTs and a hemoglobin threshold of 75 ng/mL of buffer, sensitivity and specificity were 94.1% (95% CI, 82.9% to 100.0%) and 87.5% (CI, 85.4% to 89.6%), respectively, for cancer and 67% (CI, 57.4% to 76.7%) and 91.4% (CI, 89.6% to 93.2%), respectively, for any clinically significant neoplasia.
Limitations: The fecal sampling method is standardized, but the sample size depends on fecal consistency. Some patients were tested while discontinuing aspirin and anticoagulant therapies. Study patients were at increased risk, and results might not apply to average-risk populations.
Conclusions: Quantitative immunochemical FOBT has good sensitivity and specificity for detection of clinically significant neoplasia. Test performance in screening average-risk populations is not known.