Summary : The study was conducted on 51 newly diagnosed type I diabetic children with a mean age of 8.57±4.02 ranging from 1.4 to 15.0 years with a male/female ratio (31/20). All cases were recruited from those attending the diabetic endocrine metabolic pediatric unit (DEMPU) children Hospital, Cairo University. Patients were divided into 3 groups according to age of diabetes onset (< 5 years, 5-10 years and 10-15 years) and 3 groups according to duration of the disease (< 2 months, 2-6 months and > 6 months). 47 age and sex matched control subjects were included. Screening for autoantibodies for type I diabetes was performed: Islet cell antibody (ICA) by ELISA ; Glutamic acid decarboxylase 65 antibody (GAD65A) by RIA and Tyrosine phosphatase antibody (IA-2A) by RIA. The present study demonstrates that GAD65As assays are more frequently encountered in type I diabetic patients than IA-2A and ICA values being 62.7%, 36.2% and 17.6% respectively and the prevalence of any one of the three autoantibody was 76.4% which was higher frequency than any of antibody alone. However the diagnostic sensitivity of the various antibody assays investigated tends to vary according to age at diabetes onset, both GAD65A and IA-2A being relatively higher at 5-10 years and in addition they are encountered more frequently in the group with diabetes duration of 2-6 months while ICA antibodies were less frequently encountered at these duration and age groups. The identification of antibodies to IA-2 in type I diabetic children who were GAD65A and ICA negative and vice versa suggests that testing for the three antibodies could increase the sensitivity of prediction in comparison with one marker alone. The existence of a group of type I diabetic children with negative serologic markers (21%) point out to the possibility of the presence of other islet epitopes that are not yet recognized by the currently available tests or to the possibility that autoimmunity is not the only factor in the pathogenesis of type I diabetes. Results of the present study indicate that the use of the three markers is more reliable than using one or two markers assays if active search for prediction of type I diabetes is required.