Abstract
In most colonoscopies performed for bowel symptoms, no significant lesions are found. To decrease the number of unnecessary colonoscopies, we evaluated the performance of two-sample fecal immunochemical testing (FIT) in ruling out significant lesions.Symptomatic patients referred for colonoscopy were instructed to perform two FITs from separate bowel movements prior to colonoscopy. Colonoscopy and pathology data were collected. Two-sample FIT was considered positive when FIT1 and/or FIT2 results were positive. Sensitivity and negative predictive value (NPV) for advanced neoplasia, advanced serrated polyps, and colitis were determined at different cutoff values.949 patients (median age 61 years, 50.6% male) from 10 centers were included. The highest NPVs and sensitivities were reached with two-sample FIT using the lowest limit of fecal hemoglobin detection (>1.7 µg Hb/g). For advanced neoplasia and CRC, this resulted in NPVs of 95.6% and 99.7%, and sensitivities of 71.7% and 93.9%, respectively. Sensitivity for advanced neoplasia was higher (84.6%) in patients with the alarm symptoms of rectal blood loss and/or anemia. NPV and sensitivity for inflammatory bowel disease were 99.3% and 83.3%, respectively. Concordant negative results were found for 675 patients (71.1%).Despite a high NPV, two-sample FIT still missed 28.3% of advanced neoplasia. Therefore, two-sample FIT may play a role in determining the need for colonoscopy in symptomatic patients, but it misses too many lesions to be used as the sole determinant for averting colonoscopy.