PTH-231 Pathways of epithelial leakage and zinc malabsorption in environmental enteropathy imaged by confocal endomicroscopy

Introduction Environmental enteropathy (EE) is associated with growth failure in children in low income countries, malabsorption, and impaired responses to oral vaccines. EE is characterised by morphological changes (villous blunting), increased permeation of macromolecules, microbial translocation, and impaired absorption. We set out to image the leakiness of the enteropathic small intestine using confocal laser endomicroscopy (CLE) of the duodenum following fluorescein injection, and correlate this with other markers of structure, absorption and microbial translocation. Method Fifty adults were recruited from a poor community in Lusaka where we have previously established that EE is ubiquitous, and then underwent CLE on the day after permeability (lactulose 5g, rhamnose 1g) and net zinc absorption from a 25mg test dose were measured. Between 35 and 225 CLE images per procedure were quantified for a range of cellular abnormalities and fluorescein leakage. Duodenal biopsies were assessed morphometrically and claudin-4 expression quantified. Of the 50 participants, one had pyloric stenosis and in 8 the image quality was sub-standard, so full data are available from 41 (14 men, mean age 32 y, 11 HIV seropositive). Watson scores were 1 in 4 people, 2 in 6, and 3 in 31, which contrasts markedly with published control data in healthy German adults in which nearly all were grade 1 ( P .0001). Results Leakage of fluorescein into the duodenal lumen was inevitably present when visible cell defects or microerosions were present, and was inversely associated with net absorption of zinc (r = -0.39; P = 0 .04). Direct markers of microbial translocation (LPS and LPS binding protein) were correlated (in HIV seropositives only) with gastric pH (r = 0.72; P = 0 .01) and with a serum marker of epithelial damage, fatty acid binding protein (FABP, r = −0.62; P = 0 .04). FABP was strongly correlated with leakage, especially in HIV infection (r = −0.69; P = 0 .007). Leakage was also inversely correlated with the intensity of expression of claudin-4 along the basolateral surface of the enterocyte (r = −0.49; P = 0 .01). In multiple linear regression, LPS was strongly correlated with cell shedding and GLP-2 in serum. Conclusion These data indicate that the damaged epithelium of EE leads to leakage of circulating molecules into the gut, and may explain at least one measure of absorptive defect. Pathways of translocation maybe more complex, with some translocation that cannot be explained by cell disruption. Translocation pathways are not well reflected by lactulose permeation. A clearer understanding of mechanisms will be important to reducing translocation in the enteropathic gut. Disclosure of interest None Declared.