Introduction Colonoscopy is an invasive test with associated risks and the majority performed in our NHS Trust are not associated with serious pathology. We aimed to determine whether colorectal cancer could be excluded, and colonoscopy resources better utilised, using faecal haemoglobin concentration (f-Hb) measurements to triage patients presenting in primary care with lower abdominal symptoms.

Method Patients referred from primary care to NHS Lanarkshire with colorectal symptoms, who were vetted straight to test for colonoscopy, were asked to participate. Patients were asked to complete a single faecal immunochemical test (FIT) for f-Hb measurement prior to colonoscopy. F-Hb levels were then correlated with colonoscopic and histopathological findings.

Results This was a diagnostic accuracy study.

From 909 invited patients, 507 patients completed a FIT test. Four hundred and eighty four (484) of them had a colonoscopy. Findings at colonoscopy revealed a total of 45 cases (9.3%) with significant pathology (11 colorectal cancers, 19 high-risk adenomas, 3 inflammatory bowel disease and 12 other colitides). These had a significantly higher (p < 0.0001) f-Hb than those with a normal colonoscopy (243) or findings with less clinical importance e.g. diverticular disease (196).

When a f-Hb cut-off of 10 µg Hb/g faeces is used the sensitivity was 68.9% (95% CI: 53.2–81.4) with a specificity of 80.2% (95% CI: 76.1–83.7). The negative predictive value was particularly high at 96.2% (95% CI: 93.5–97.8) with a positive predictive value of 26.3% (95% CI: 18.8–35.3). Of the 243 with normal colonoscopies, 81.2% had f-Hb <10 µg Hb/g faeces.

Conclusion Considerable potential exists with using f-Hb measurements to assist with triaging symptomatic patients within in a clinical pathway, to either colonoscopy or clinic assessment. This is particularly true with the high negative predictive value of f-Hb suggesting that it can rule out colorectal cancer and other significant colorectal disease with a negative test. No cases of colorectal cancer, in our population, would be missed when using a f-Hb level of <10 µg Hb/g faeces as a cut off. Those with a positive f-Hb could be fast-tracked as they are at higher risk of significant colorectal disease while those with a lower f-Hb level might be best served by initial clinic review. This would help to conserve hospital resources, and ensure timely and appropriate assessment for symptomatic patients.

Disclosure of interest None Declared.