Residual insulin secretion in individuals with type 1 diabetes in Finland: longitudinal and cross-sectional analyses

Background Contrary to the presumption that type 1 diabetes leads to an absolute insulin deficiency, many individuals with type 1 diabetes have circulating C-peptide years after the diagnosis. We studied factors affecting random serum C-peptide concentration in individuals with type 1 diabetes and the association with diabetic complications.Methods Our longitudinal analysis included individuals newly diagnosed with type 1 diabetes from Helsinki University Hospital (Helsinki, Finland) with repeated random serum C-peptide and concomitant glucose measurements from within 3 months of diagnosis and at least once later. The long-term cross-sectional analysis included data from participants from 57 centres in Finland who had type 1 diabetes diagnosed after 5 years of age, initiation of insulin treatment within 1 year from diagnosis, and a C-peptide concentration of less than 10 nmol/L (FinnDiane study) and patients with type 1 diabetes from the DIREVA study. We tested the association of random serum C-peptide concentrations and polygenic risk scores with one-way ANOVA, and association of random serum C-peptide concentrations, polygenic risk scores, and clinical factors with logistic regression.Findings The longitudinal analysis included 847 participants younger than 16 years and 110 aged 16 years or older. In the longitudinal analysis, age at diagnosis strongly correlated with the decline in C-peptide secretion. The cross-sectional analysis included 3984 participants from FinnDiane and 645 from DIREVA. In the cross-sectional analysis, at a median duration of 216 years (IQR 125-312), 776 (194%) of 3984 FinnDiane participants had residual random serum C-peptide secretion (>002 nmol/L), which was associated with lower type 1 diabetes polygenic risk compared with participants without random serum C-peptide (p<00001). Random serum C-peptide was inversely associated with hypertension, HbA(1c), and cholesterol, but also independently with microvascular complications (adjusted OR 061 [95% CI 038-096], p=0033, for nephropathy; 055 [034-089], p=0014, for retinopathy).Interpretation Although children with multiple autoantibodies and HLA risk genotypes progressed to absolute insulin deficiency rapidly, many adolescents and adults had residual random serum C-peptide decades after the diagnosis. Polygenic risk of type 1 and type 2 diabetes affected residual random serum C-peptide. Even low residual random serum C-peptide concentrations seemed to be associated with a beneficial complications profile.