Re: transient marked decrease in serum transaminases in chronic hepatitis C after surgery

TO THE EDITOR: I thank Dr. Radaelli and colleagues for their interest and for the opportunity to comment further on our results. We did emphasize that these results were obtained from an area with a high background prevalence of celiac disease (CD) and may not be applicable elsewhere. Most recent estimates suggest a CD prevalence in Northern Ireland as high as 1 in 120 (1). We agree, however, that selection of patients for endoscopy by referring doctors may also partly explain the high prevalence within the group studied. Primary care practitioners often prescribe H licobactereradication therapy after positive serological testing or empirical proton pump inhibitors, reserving endoscopy referral for treatment failures. Despite guidelines, we have evidence that these strategies are being increasingly applied to patients of all ages. The etiology of celiac dyspepsia is uncertain, but it seems to be partly due to dysmotility (2). It might well respond poorly toHelicobacter pylorieradication or acid suppression, and would therefore be over-represented among patients referred for endoscopy as treatment failures. In our earlier study of endoscopic markers of villous atrophy (VA) among patients having open access endoscopy, the prevalence was 1:63, but referral guidelines were designed to optimize the yield of peptic disease (3). Different referral patterns may explain the difference between our results and those of Radaelli et al. I reviewed 150 consecutive patients, a year after the study series, to determine the prevalence of reported endoscopic markers outside the confines of prospective research. As before, patients undergoing endoscopy for the specific purpose of duodenal biopsy were excluded. Nineteen had anemia without gastrointestinal (GI) symptoms: none of these had endoscopic markers of celiac disease or villous atrophy on biopsy. Of the remaining 131, who had endoscopy for dyspeptic symptoms, four had endoscopic markers. All four had biopsy-confirmed VA, giving a prevalence of 1:33. We emphasized that reliance on endoscopic markers for biopsy selection will result in underdiagnosis of CD. However, the alternative is to obtain duodenal biopsies in every patient undergoing endoscopy for upper GI symptoms, which would have major cost implications. A policy of biopsy in these patients in whom there are endoscopic markers, or where some aspect is suggestive of CD ( i.e., anemia, weight loss, flatulence, diarrhea, or family history) should optimize the opportunity for diagnosis of CD during routine duodenoscopy.