Gender differences in faecal haemoglobin concentration.

Abstract

In his valuable summary, Professor Fraser notes that the distribution of haemoglobin levels within a population may be useful in matching positivity thresholds to regional health service capacity. We further explored the effect of gender on faecal haemoglobin levels by analysing these concentrations in single sample faecal immunochemical tests (FIT, OC Sensor, Eiken Chemical Company, Japan) completed within two weeks prior to scheduled colonoscopy (for any indication, n ¼ 2009). Gender differences were similar to those in our recent publication. 1 We therefore investigated: (1) the baseline faecal haemoglobin level of males and females who had been found at colonoscopy to have a normal colon; and (2) the sensitivity of the same haemoglobin cut-off in males and females to detect advanced neoplasia (colorectal cancer or advanced adenoma). Among people with no colorectal disease at colonoscopy, a statistically significant difference between sexes in faecal haemoglobin concentration was evident. The median level in males was 1.0mg Hb/g faeces (25th–75th percentile: 0 – 3.2mg Hb/g faeces, n ¼ 179) and in females 0.6mg Hb/g faeces (25th–75th percentile: 0 – 1.4mg Hb/g faeces, n ¼ 219; p < 0.001). We then assessed whether these differences would also exist in cases of advanced neoplasia, among 234 males (21% of colonoscopy findings) and 150 (14%) females. Using the positivity cut-off of 10mg Hb/g faeces in a FIT, 50% (118/234) of males, but only 43% (64/150) of females were positive. To enable detection of the same proportion of advanced neoplasia in females in our study would require the positivity cut-off of 5.4mg Hb/g. The difference in sensitivity (50% vs 43%) is not statistically significant in this population (Chi-squared test p ¼ 0.14), but the same positivity cut-off, applied to nation-wide screening programmes, may lead to more missed cancers and adenomas in females. This is a potential factor in the higher rate of female FIT-negative interval cancers. 2 When establishing positivity thresholds for quantitative FIT in screening programmes, differing faecal haemoglobin levels between males and females, or between different age bands, influence the effectiveness of screening. Furthermore, regional differences may necessitate different thresholds to achieve equivalent advanced neoplasia detection rates. However, decreasing test positivity cut-off, while increasing neoplasia sensitivity, can also increase false positive rate. Further work is needed to establish the best screening strategy to maximise neoplasia detection in all groups of the community to improve equity of screening, while taking into consideration colonoscopy capacity.