Several workshops attempting to standardize the islet-cell antibody (ICA) assay have demonstrated marked but unexplained interlaboratory variation. In this study, the effect of using 10 different pancreata to quantitate ICA in a standardized two-color immunofluorescence test within the same laboratory was evaluated. Sera from 11 recent-onset insulin-dependent diabetic patients and one healthy control were analyzed. The negative control serum was recorded as ICA negative in all 10 pancreata. The ICA end-point titers varied, with 1.1 -2.4 titration steps in standard deviations in the 10 different pancreata. The interassay variability within the same pancreas was 0.6-0.9 titration steps and therefore could not alone explain the interpancreatic variation. The end-point titers in the different pancreata were highly correlated (rs = 0.71l.O), supporting the hypothesis that the pancreatic tissue strongly influences the ICA levels. A quality index was developed for selection of pancreata with qualities for ICA analyses. The international ICA standard serum used to express the results in Juvenile Diabetes Foundation units diminished the interpancreatic variability.