Frederik Willem Janssen is head of the Chemistry Department, Food Inspection Service in Zutphen, and a subsidiary of the Inspector of Health Protection (similar to the FDA in America). Their lab has a special interest in.... modified gluten, edible packaging materials (which may contain gluten), and detection of hidden gluten in foods, including the development of improved detection methods. He is also a member of the Medical/Scientific Advisory Committee of the Dutch Celiac Society.
Distillation quite effectively removes the gluten and it is very unlikely that splashes of fermented (we call it "moutwijn", i.e. malt wine, can't remember the correct English word for it) will be carried over to the final distillate. If they are present they must have been added afterwards. A couple of years ago we analyzed some distilled liquor for presence of gluten proteins but we couldn't detect any in this set (about 40 different types). The beer test, which consisted of a set of 50 different brands, showed that most brands (35) did contain immunoreactive protein in amounts between 1 and 200 mg/liter. Only 15 contained less than 1 mg/liter. There was a strong correlation between the gluten content and whether wheat had been used as an ingredient!
They did find gluten in distilled liquor! The levels varied from zero to 200-mg gluten/liter. The highest amount was found in a "Creme de Framboise" (200 mg/liter) but second was a French brandy VSOP with a score of 180 m g/liter. A Dutch gin was negative, which might be an indication that gluten in these type of liquor is not a carry over to the distillate! My guess is that this gluten is derived from the caramel coloring, though there is no proof about this yet. I always advise sensitive patients to abstain from brown colored liquor!
I would like to stress that the determination of gluten in these types of products is very unreliable and we have to count with false positive as well as false negative values. The gluten proteins could have been broken down to small (but still toxic) peptides and in that case a sandwich-type ELISA might produce false negative results because in that case you always need to two epitopes (binding sites for the antiserum) on one molecule to get a positive reaction. A competitive type assay would be the choice for this type of product, but we haven't tried this type of analysis yet. We did use it on a soy sauce, which was prepared with wheat gluten and didn't find any gliadin, which might be an indication that gluten had been broken down to very small peptides with less than one binding site.
Frederik Willem Janssen, Zutpen, The Netherlands.