Coeliac biopsies: numbers are valid, alphabets not

We thank Gut for requesting responses to this letter from Oberhuber and colleagues.1 We were amazed that they chose to dismiss our intraepithelial lymphocytes (IEL) data as invalid2 on grounds of not being age/gender matched with controls. But this disregards all the other data of clinicians and pathologists in the field, and the rigorous reviewing processes that our article went through before publication in Gut. This dismissal reflects an aberrant viewpoint1 lacking widespread support. In real life, pathologists do sign off biopsies, while diagnoses are rendered without full clinical details, although with appropriately worded caveats.

We aimed, specifically, to reduce the enormous overlap between control/coeliac IEL counts, by taking the lowest (receiver operating characteristic) value favouring controls (table 1). Incidentally, female cohort size between our groups only differed by ~10%, while their age ranges (44 vs 37: P (mean±SD)=ns) were almost identical. In effect, statistically, the...

http://ift.tt/2jaY6vV