Abstract
The source of variability in the outcome of islet cell antibody (ICA) determinations due to pancreas and serum were independently assessed with a quantitative assay. The effect of choice of pancreatic substrate and serum specimen was assessed by testing for parallelism among twenty-five dilution curves of five sera from diabetic subjects and five pancreata. The five by five format allowed for five tests of pancreatic variability (one serum against five pancreata) and five tests of serum variability (one pancreas against five sera). Analyses of variance and covariance of these dilution curves indicate that test outcome is affected by both the sera (5 tests, p values 10-6 to 10-7) and the pancreas (5 tests, p values 10-3 to 10-5). The data support heterogeneity among islet cell antibody isotypes and raise the possibility of diversity of islet cell antigens. Antigenic diversity of pancreatic substrates may interfere with inter- and intra-laboratory comparisons of islet cell antibody determinations. The strong effect of pancreas type on ICA outcome infers possible heterogeneity of the islet cell antigen(s) capable of reacting with human islet cell antibodies.