Using the faecal immunochemical test in patients with rectal bleeding: evidence from the NICE FIT study.

Abstract

The aim of this work was to investigate whether the faecal immunochemical test (FIT) could safely rule out colorectal cancer (CRC) in patients with rectal bleeding (RB). This was a multicentre, double-blinded diagnostic accuracy study in 50 National Health Service hospitals. Patients referred from primary care with suspected CRC on an urgent 2-week-wait pathway were asked to perform a FIT prior to colonoscopy. The primary outcome measure was the sensitivity of the FIT for CRC in patients with RB versus nonrectal bleeding symptoms (NRB). The secondary outcome measures included the diagnostic accuracy of the FIT for CRC and other serious bowel disease. Of 9822 patients included in the study, 3143 (32.0%) were referred with RB. CRC was present in 4.7% of patients with RB versus 2.7% of patients with NRB (p < 0.05). Faecal haemoglobin (f-Hb) was detectable (>2 µg/g) in 44.1% of patients with RB and 33.9% with NRB (p < 0.05). In RB patients, CRC was present in 10.4% when f-Hb was >2 µg/g compared with 0.1% when f-Hb was not detected. Flexible sigmoidoscopy in this group would further reduce the risk of CRC to 0.03%. The sensitivity of the FIT for CRC in RB versus NRB groups was 98.6% (95% CI 95.2%-99.8%) vs 95.6% (91.5%-98.1%) for f-Hb >2 µg/g and 96.6% (92.2%-98.9%) vs 86.3 (80.4%-90.9%) for f-Hb >10 µg/g. Faecal haemoglobin is not always detectable in patients with RB; 56% of patients had undetectable f-Hb (<2 µg/g) and CRC was present in 0.1%. The high sensitivity of the FIT can be used to rule out CRC in patients with RB and triage them more appropriately for investigation.