Letter: Adnab-9 as a potential non-invasive biomarker for prediction of malignancy in coeliac disease



We read with great interest the article by Lebwohl et al.[1] regarding the need for objective biomarkers to predict morbidity and mortality in coeliac disease (CD). Villous atrophy is a hallmark of CD, and it has been postulated that defective mucosal detoxification of carcinogens in atrophic villi leads to the observed increased incidence of cancer in patients with poorly controlled CD.[2]

While the authors found that persistent villous atrophy was not a suitable prognostic biomarker for mortality in CD, they nonetheless reported a trend (= 0.2542) towards increased mortality from malignancies.

The Adnab-9 antibody was developed to detect premalignant lesions in the gastrointestinal tract,[3] and has been validated in various studies.[4-6] To test whether Adnab-9 can be used to estimate risk of developing malignancy in patients with CD, we measured levels of Adnab-9 antigen in stool supernatants from a small cohort of CD patients on normal or gluten-free diets (GFD) (Figure 1).

Figure 1.

Adnab-9 immunoreactivity of faecal supernatants from coeliac disease patients on normal or gluten-free diet (GFD). Adnab-9 binding was detected using an ELISA performed on the faecal supernatants obtained from 27 adult CD patients from the greater area of Rome, Italy, who were either on normal diets (5M/17F, mean age 40.7, range: 18–71), or GFD (5F, mean age 33.0, range: 24–52). All subjects' serum and faecal supernatant samples were endomysial antibody positive. Experiments were performed in an independent laboratory blinded to the treatment of the patients from which stool samples were obtained. Results are expressed in optical density(OD)/5 μg protein minus background. Horizontal bars designate means. Note that all patients in the GFD group except one have very low Adnab-9 binding. Mean Adnab-9 binding is significantly reduced in the GFD group (P = 0.045). Subanalyses considering only female or age-matched normal-diet patients yielded comparable results (17F, mean age 40.8: mean OD 1.25; 13F/2M, mean age 31.9: mean OD 1.34). Statistical analysis was performed using unpaired t-test with Welch correction.

The results show that while patients on normal diets have widely variable levels of Adnab-9 antigen in stool water, all patients on GFD but one had very low levels of Adnab-9 binding. This decrease parallels the greatly diminished risk for developing GI malignancies such as oesophageal cancer or small bowel adenocarcinoma in patients with GFD.[7] Together, this suggests that Adnab-9 might be a novel marker to predict risk of malignancy in patients with CD.

In summary, there is clear need for objective biomarkers to predict cancer risk in patients with CD. Our results suggest that Adnab-9 immunoreactivity might serve this purpose. The observed decrease in Adnab-9 binding with GFD was similar to previously observed decrease in anti-tissue transglutaminase antibodies.[8] However, as Adnab-9 is a premalignant marker, we hypothesise that Adnab-9 detection in stool water may be a more accurate and specific means to detect patients with CD who are at high risk to develop small bowel cancers. Prospective studies are needed to further investigate these ideas.


The opinions expressed in this publication are not necessarily those of the VHA or the US Federal Government. Declaration of personal and funding interests: None.