CONTEXT: Insulin-binding antibodies may produce severe dysglycaemia in insulin-naïve patients ('insulin autoimmune syndrome' (IAS) or Hirata disease), while rendering routine insulin assays unreliable. OBJECTIVE: To assess the performance of clinically used insulin assays and an optimal analytical approach in the context of IAS. DESIGN: Observational biochemical study of selected patients with hyperinsulinaemic hypoglycaemia. PATIENTS: Three patients without diabetes with recurrent spontaneous hyperinsulinaemic hypoglycaemia and 'positive' insulin antibodies. MEASUREMENTS: A panel of clinically used insulin assays (Siemens ADVIA® Centaur, Siemens Immulite® 2000, DiaSorin LIAISON® XL, PE AutoDELFIA® and the Beckman Coulter Access® 2) were used before and after plasma dilution or polyethylene glycol (PEG) precipitation. Anti-insulin IgG antibodies were measured by Isletest™ -IAA ELISA. Gel filtration chromatography (GFC) was undertaken with and without preincubation of plasma with exogenous insulin. RESULTS: Dilution of IAS plasma with assay-specific buffer increased insulin recovery, supporting negative immunoassay interference by antibodies. PEG precipitation of IAS plasma decreased insulin recovery using all assays except the Immulite® 2000. GFC discriminated high molecular weight and monomeric insulin, while ex vivo addition of exogenous insulin to plasma increased insulin bound to antibody, thereby improving the sensitivity of detection of insulin immunocomplexes. CONCLUSIONS: Immunoprecipitation with PEG must be used with caution in screening for insulin-antibody complexes as results are assay dependent. GFC with addition of exogenous insulin can identify significant insulin immunocomplexes with enhanced sensitivity, with attendant greater clinical utility and avoidance of radiolabelled reagents.