The transgastric approach is a novel method for obtaining liver biopsies in patients undergoing upper gastrointestinal endosonography. Avoidance of vascular puncture and ability to acquire tissue in patients with obesity or ascites offers a practice niche for this technique. Although several series have reported on specimen adequacy, biopsy core length, and number of portal tracts, none has addressed the diagnostic challenges presented by the fragmented nature of these specimens. We systematically evaluated 113 transgastric liver biopsies obtained for diagnosis of parenchymal liver disease by 3 needle types and compared them with 100 percutaneous and 100 transjugular liver biopsies, respectively. Parameters recorded were number of tissue cores, sizes of longest and shortest cores, numbers of complete and incomplete portal tracts, morphologic characteristics, and adequacy of specimen for diagnosis and staging. In contrast to percutaneous and transjugular liver biopsies, transgastric biopsies often contained >10 tissue fragments and smaller tissue cores. In addition, 2 of the 3 types of transgastric needles obtained less numbers of complete portal tracts. Transjugular biopsies were also smaller and contained less number of complete portal tracts than percutaneous specimens but, unlike transgastric biopsies, only rarely contained >10 tissue fragments. Specimen adequacy for diagnosis and staging was 80%, 100%, and 98% for transgastric, percutaneous, and transjugular biopsies, respectively. Difference in specimen adequacy is related to tissue fragmentation of transgastric liver biopsies rather than biopsy core length or numbers of complete portal tracts. Tissue fragmentation is particularly challenging for staging chronic liver disease.