Objective: To measure serum levels of immunoglobulin A byimmunoenzymatic assay (ELISA) in type 1 diabetes mellitus (DM-1)patients and to verify the prevalence of immunoglobulin A deficiency(IgAD) in diabetic patients. Methods: The serum immunoglobulin Alevel was determined in 149 DM-1 patients by three methods. IgADwas defined as serum immunoglobulin A level lower than 5 mg/dl.If serum immunoglobulin A level was undetectable by turbidimetry,radial immunodiffusion was performed in low plate concentration.For patients with undetectable serum immunoglobulin A levelby the two previous methods, quantification was performed byELISA. In patients with IgAD, the levels of immunoglobulins Gand M were measured by turbidimetry to exclude other humoralimmunodeficiencies. Results: Out of 149 DM-1 patients evaluated,141 (94.6%) had normal serum immunoglobulin A levels byturbidimetry. Eight patients (5.3%) had undetectable serumimmunoglobulin A levels by turbidimetry and radial immunodiffusion.In these eight patients, the determination of serum immunoglobulinA was performed by ELISA, a more sensitive method. Very lowlevels of serum immunoglobulin A were detected in these diabeticpatients. In all diabetic patients, immunoglobulins G and M werenormal for age by turbidimetry. All 150 patients of the Control Grouphad normal serum immunoglobulin A levels by ELISA. Conclusions:There was a significantly higher prevalence of immunoglobulindeficiency among DM-1 patients (5.3%). Measurement of serumimmunoglobulin A is necessary in all DM-1 particularly before someimmunoglobulin A antibody screening. Patients with IgAD may havefalse-negative results for celiac disease screening tests involvingimmunoglobulin A antiendomysium and antigliadin antibodies.