Pediatr Diabetes. 2011 May;12(3 Pt 1):142-9. doi: 10.1111/j.1399- Persistently autoantibody negative (PAN) type 1 diabetes mellitus in children. Hameed S, Ellard S, Woodhead HJ, Neville KA, Walker JL, Craig ME, Armstrong T, Yu L, Eisenbarth GS, Hattersley AT, Verge CF. SourceEndocrinology, Sydney Children's Hospital, Randwick, NSW, Australia School of Women's and Children's Health, University of New South Wales, Sydney, Australia Peninsula Medical School, Exeter, UK St George Hospital, Kogarah, Australia Barbara Davis Center for Childhood Diabetes, Aurora, CO, USA. Abstract Hameed S, Ellard S, Woodhead HJ, Neville KA, Walker JL, Craig ME, Armstrong T, Yu L, Eisenbarth GS, Hattersley AT, Verge CF. Persistently autoantibody negative (PAN) type 1 diabetes mellitus in children. Background: Autoantibody-negative children diagnosed with type 1 diabetes might have unrecognized monogenic or type 2 diabetes. Research design and methods: At diagnosis of type 1 diabetes (between ages 0.5 and 16.3 yr, n = 470), autoantibodies [glutamic acid decarboxylase (GAD), insulinoma-associated protein 2 (IA2), insulin autoantibodies (IAA), and/or islet cell antibody (ICA)] were positive (ab+) in 330 and negative in 37 (unknown in 103). Autoantibody-negative patients were retested at median diabetes duration of 3.2 yr (range 0.9-16.2) for autoantibodies (GAD, IA2, ZnT8), human leukocyte antigen (HLA) typing, non-fasting C-peptide, and sequencing of HNF4A, HNF1A, KCNJ11, and INS. Results: Nineteen (5% of 367) remained persistently autoantibody negative (PAN), 17 were positive on repeat testing (PORT), and 1 refused retesting. No mutations were found in PORT. One PAN was heterozygous for P112L mutation in HNF1A and transferred from insulin to oral gliclazide. Another PAN transferred to metformin and the diagnosis was revised to type 2 diabetes. The remaining 17 PAN were indistinguishable from the ab+ group by clinical characteristics. HLA genotype was at high risk for type 1 diabetes in 82% of remaining PAN and 100% of PORT. After excluding patients with diabetes duration <1 yr, C-peptide was detectable more frequently in the remaining PAN (7/16) and PORT (6/17) than in a random selection of ab+ (3/28, p = 0.03). Conclusions: The diagnosis of type 1 diabetes should be reevaluated in PAN patients, because a subset has monogenic or type 2 diabetes. The remaining PAN have relatively preserved C-peptide compared with ab+, suggesting slower β-cell destruction, but a very high frequency of diabetogenic HLA, implying that type 1B (idiopathic) diabetes is rare.