As a consequence of its large volume, a microencapsulated islet graft can be implanted only into the peritoneal cavity. The graft volume can be reduced by using small capsules. However, reduction of the diameter of the capsules holds a certain risk, because with smaller capsules, more islets may be found to protrude from the capsules. We have developed a lectin binding assay which, after encapsulation, specifically labels islets or parts of islets that are insufficiently immunoprotected as a consequence of inadequate, and particularly incomplete, encapsulation. With this assay, we found that a reduction of the capsule diameter from 800 micrometers to 500 micrometers was associated with an increase in the percentage of inadequately encapsulated islets from 6.3+/-1.2% to 24.2+/-1.5%. The in vivo significance of this finding was investigated by performing allotransplantations with large diameter (700-800 micrometers) and small diameter (400-500 micrometers) capsules. With large-capsule islet grafts, all recipients (n=5) became normoglycemic for 7-16 weeks, whereas with small-capsule islet grafts, only one of seven recipients became normoglycemic. The in vivo significance of inadequate encapsulation was further substantiated by our finding that most large capsules were floating freely in the peritoneal cavity without any cell adhesion, whereas the vast majority of small capsules was found to be adherent to the surface of intra-abdominal organs and infiltrated by immune cell elements characteristic of both an allograft reaction and a foreign body reaction. We conclude that successful use of capsules with small diameters requires further study to determine which factors in the encapsulation procedure should be modified to reduce the number of inadequate small capsules.