The Effect of the Variability in Fecal Immunochemical Test Sample Collection Technique on Clinical Performance

• Published in: Cancer epidemiology, biomarkers & prevention biomarkers & prevention, 2021-01, Vol.30 (1), p.175-181
• Main Authors: Symonds, Erin L, Fraser, Callum G, Bastin, Dawn, Berwald, Grace, Young, Graeme P
• Format: Article
• Language: English
• ISSN: 1055-9965; 1538-7755
• DOI: 10.1158/1055-9965.EPI-20-0984

Fecal immunochemical test (FIT) performance can be affected by post-collection variables. Collection technique might also affect fecal hemoglobin concentration (f-Hb). Variation in quantity of feces collected in samples returned in a colorectal cancer detection program, and the effects of under-sampling, were assessed. Collection devices obtained from patients undergoing FIT were assessed for the color (in five classes) of the feces in buffer, mass, and f-Hb. Associations between these were examined in an study on Hb-spiked feces. Variables possibly associated with under-sampling were investigated using multivariable logistic regression. The effect of low sample mass on clinical performance (false-negative results) was determined. Of 6,898 samples collected by 3,449 individuals (46.9% male, median age: 65.3 years), the buffer was lightest in color in 362 (5.2%), and darkest in 420 (6.1%). Samples with the lightest color had a significantly lower f-Hb compared with all darker classes ( < 0.001). Mass was recorded for 650 devices: The lightest colored samples had significantly lower mass ( < 0.05). The correlation between mass and f-Hb was confirmed ( = 0.897, < 0.001). Low mass was not associated with age, sex, or technical factors ( > 0.05). Under-sampling related to the lightest color was not associated with false-negative results for colorectal cancer and advanced adenoma, but was for all neoplasia and inflammatory bowel disease. Wide variation existed in the amount of feces collected. Under-sampling results in lower measured f-Hb and may increase false-negative results. Color of sample buffer could be used to identify inadequate sampling.