A series of 148 diabetic patients were studied for the presence of soluble immune complexes using five different screening techniques. The percentage of positive results was 26% with direct nephelometry and PEG-C4, 27% with PEG-IgG, 33% with radiolabelled Clq binding and 57% with a specific technique for detection of insulin-anti-insulin immune complexes. The percentages of positivity in a group of 40 healthy donors were 2.5% for direct nephelometry and radiolabelled Clq binding, 5% for the PEG-C4 technique, and 10% for the PEG-IgG technique. Sixteen percent of the patients studied had positive results in three or more of the screening tests. When the results of the different screening tests in all patients and controls were compared among themselves, we found correlation coefficients between-0.01 (p = 0.854) when the direct nephelometry and the PEG-C4 tests were compared and 0.29 (p < 0.0003) when the direct nephelometry and PEG-IgG tests were compared. When the results of each test for the whole group of patients and the group of normal healthy donors were compared, significant differences were found for direct nephelometry (p = 0.004), PEG-IgG, PEG-C4, and insulin-anti-insulin immune complexes (p < 0.0001), as well as for anti-insulin antibodies (p < 0.001); no significant difference was observed when the results of radiolabelled Clq binding in diabetics and controls were compared (p = 0.2). Significant correlations were found between the results of several screening tests for soluble immune complexes, insulin dosage, and clinical or biochemical expressions of microangiopathy, nephropathy, or vasculopathy. These correlations were more consistent when we divided the patients into normal or abnormal groups for proteinuria, microangiopathy, and diabetic complications and considered the number of positive tests in each patient rather than the results of individual tests.