Abstract
Gestational diabetes mellitus (GDM) is increasingly recognized as a heterogeneous metabolic disorder comprising pathophysiological phenotypes that differ in their underlying mechanisms, clinical presentation and long-term outcomes. While pregnancy-induced insulin resistance remains the predominant pathogenic mechanism, a subset of women develops hyperglycemia largely because of reduced pancreatic β-cell reserve accompanied by islet autoimmunity. This autoimmune phenotype links gestational diabetes with autoimmune diabetes, but its boundaries and clinical significance remain incompletely defined. We aimed to review current knowledge of the autoimmune phenotype of GDM, with focus on immune pathophysiological mechanisms, biomarkers, clinical characteristics and implications for precision medicine, and to distinguish findings demonstrated directly in GDM from mechanisms extrapolated from type 1 diabetes (T1D). This narrative review was based on a structured PubMed/MEDLINE search from database inception through the end of October 2026, supplemented by targeted searches, manual screening of reference lists, and relevant clinical guidelines and consensus documents. Evidence obtained directly in women with GDM was considered alongside selected mechanistic literature from T1D and adult autoimmune diabetes. In some women with GDM, islet autoantibodies are associated with impaired β-cell function and an increased risk of autoimmune diabetes after delivery. The cellular mechanisms underlying these associations remain insufficiently studied in GDM. Pregnancy may act as a physiological metabolic stress test that exposes previously compensated β-cell dysfunction associated with islet autoimmunity, although its influence on the underlying autoimmune process remains uncertain. Islet autoantibodies are enriched among women with GDM, but their frequency varies widely between populations and assays and is lower after delivery than during pregnancy. The number and persistence of antibodies, particularly when combined with reduced endogenous insulin secretion, carry considerably more prognostic weight than a single positive result. Cellular immune mechanisms are well characterized in T1D but remain largely unvalidated in GDM. Recognition of the autoimmune phenotype extends the concept of GDM beyond a purely glucose-based classification. It is best viewed as a spectrum ranging from isolated seropositivity to pregnancy-unmasked autoimmune diabetes rather than as a uniform independent entity. A phenotype-oriented approach combining clinical features, autoantibody profiles and measures of β-cell function may facilitate earlier recognition of women at risk and more individualized follow-up, but universal autoantibody screening and antibody-specific treatment are not currently supported.