OBJECTIVE To evaluate the prevalence of IgG antibodies to bovine serum albumin (BSA) in a cohort of Brazilian children and young adults with IDDM. RESEARCH DESIGN AND METHODS Sera from 81 subjects with < 1 year of IDDM (group 1), III subjects with > 1 year of IDDM (group 2), and 207 normoglycemic subjects were tested using an immunofluorimetric assay. A receiver-operating-characteristic curve was used to establish the threshold of anti-BSA antibody titers defining the positivity of the assay. RESULTS The distribution of the fluorimetric index (FI) of anti-BSA antibodies did not have a gaussian profile. Rank sum of FI was significantly higher in patients than in control subjects (P < 0.0001). Average logFI values of both IDDM groups were significantly higher than that of the control group (P < 0.005 for both groups). There was a trend toward higher FI levels in group 1 than in group 2 (P = 0.06). A FI cutoff of 0.7 optimized the ratio of true-positive to false-positive of the assay, with the best equilibrium between sensitivity and specificity. The prevalence of anti-BSA antibodies was 52% in group 1, 47% in group 2, and 28% in the control group (P = 0.0001). An independent association between anti-BSA antibodies and IDDM, with an odds ratio of 3.03 (P < 0.0001), was observed in a logistic regression analysis. However anti-BSA antibodies explained only 5% of the variability of IDDM versus NIDDM. CONCLUSIONS Our results confirm that the prevalence of anti-BSA antibodies is higher in IDDM subjects than in control subjects, even after 1 year of diabetes. However, a large overlap of antibody titers is observed in patients and control subjects, suggesting that anti-BSA antibodies are neither sensitive nor specific markers of IDDM.