Faecal immunochemical testing (FIT): Sources of analytical variation based on three years of routine testing in the context of DG30

Aims To determine analytical capabilities of a commonly used faecal immunochemical test (FIT) to detect haemoglobin (Hb) in the context of NICE guidance DG30, and the likely use of FIT to reprioritise patients delayed by the COVID-19 pandemic. Methods Data obtained from independent verification studies and clinical testing of the HM-JACKarc FIT method in routine primary care practice were analysed to derive analytical performance characteristics. Results Detection capabilities for the FIT method were 0.5 µg/g (limit of blank), 1.1 (limit of detection) and 15.0 µg/g (limit of quantification). 31 of 33 (94%) non-homogenised specimens analysed in triplicate were consistently categorised relative to 10 µg/g compared to all 33 (100%) homogenised specimens. Imprecision in non-homogenised specimens was higher (median 27.8%, (range 20.5% - 48.6%)) than in homogenised specimens (10.2%, (7.0 to 13.5%)). Considerable variation was observed in sequential clinical specimens from individual patients but no positive or negative trend in specimen degradation was observed (p=0.26). Conclusions The FIT method is capable of detecting Hb at concentrations well below the DG30 threshold of 10 µg/g. However, total imprecision is considerable when including sampling variation. Binary categorisation against a single defined threshold above and below 10 µg/g was more consistent and improved following specimen homogenisation. This approach may be more appropriate when reporting results for symptomatic patients tested in primary care, including those who have had definitive investigation delayed by the COVID-19 pandemic and need to be re-prioritised. Key Messages Faecal immunochemical testing (FIT) is increasingly used to detect blood at low haemoglobin (Hb) concentrations in specimens from symptomatic primary care patients but the analytical characteristics in this context have not been fully documented. A commonly used FIT method showed good capability in a routine UK clinical setting to detect Hb at the NICE recommended threshold of 10µg/g. Imprecision estimates were considerable when sampling was considered, even above the limit of quantification of 15 µg/g. Analytical variability appears too high for reliable reporting of quantitative Hb concentrations: reporting positive or negative results around a threshold of 10µg/g appears more appropriate after sample homogenisation. Dichotomous FIT reporting is likely to be an important tool to risk stratify patients with lower GI cancer symptoms who have had their test deferred due to the COVID-19 pandemic