P0390 VARIABILITY OF HISTOLOGIC LESIONS IN RELATION TO BIOPSY SITE IN CELIAC DISEASE

Abstract

Introduction: Histologic lesions of celiac disease (CD) range from increased intraepithelial lymphocytes (IEL) to crypt hyperplasia (CH) to villous atrophy (VA). It is believed that the optimal biopsy site for detection of VA is the distal duodenum or proximal jejunum and lesions may have a patchy distribution which could lead to misdiagnosis. We wanted to verify these concepts and establish whether biopsies should always be taken from these areas in order to make a diagnosis of CD. Methods: One-hundred and 12 consecutive children with positive anti-endomysium (EMA) or anti-tissue transglutaminase (tTGA) antibodies who had been referred for suspected CD were enrolled in a prospective fashion. All patients underwent upper GI endoscopy where 4–5 biopsies were taken sequentially with standard biopsy forceps from: Treitz and/or distal duodenum (T-D3), intermediate duodenum (D2), proximal duodenum (D1) and duodenal bulb (B). Biopsies were oriented, fixed in formalin, stained with H&E and subjected to immunohistochemistry with anti-CD3 antibodies for IEL count. Histologic lesions were classified according to Marsh criteria modified by Oberhuber. Results: One-hundred and 10 of 112 patients, all HLA DQ2 or DQ8 positive, had a final diagnosis of CD (59 classic, 28 atypical, 23 silent). Of these 110, 102 (92.7%) had VA with lesion type 3: a) mild, b) moderate or c) severe in at least one site and 94/110 (85.4%) had VA in all sites though with variable degree. VA of identical degree was present in all biopsy sites in 55/110 (50%) patients. Severe VA (type 3c) was present in at least one site in 85/110 (75%), in all sites in 50/110 (45.4%) and was distributed as follows: B 50/110 (45.4%), D1 48/110 (43.6%), D2 70/110 (63.6%), D3/T 83/110 (75.4%), with a significant increase in aborad direction (chi2>26.22 with a=0.01 and g.l.=12). Eight/110 (7.2%) CD patients had exclusively type 1 or 2 lesions (increased IEL ± CH), 9/110 (8.1%) had lesion variability = 1 degree (e.g. type 1 and 2, type 2 and 3a) but no patient had lesion variability >1 degree and none had normal biopsies. There was no statistically significant correlation between type or distribution of histologic lesions and clinical presentation of CD, though interestingly all patients with silent CD had type 3c lesions throughout the duodenum. Conclusion: Mucosal atrophy is present in 85% of CD patients and TVA is significantly more frequent in more distal areas (D3/T). However, in half of the patients VA has the same degree throughout the duodenum including B, and no duodenal areas are histologically normal. Therefore, in genetically susceptible children with positive serology, a diagnosis of CD can reliably be made aven if biopsies are not taken from the distal duodenum or jejunum.