Predictive biomarkers of rapidly developing insulin deficiency in children with type 1 diabetes.

Abstract

Introduction: The rate of progression to complete insulin deficiency varies greatly in type 1 diabetes. This constitutes a challenge, especially when randomizing patients in intervention trials aiming to preserve beta cell function. This study aimed to identify biomarkers predictive of either a rapid or slow disease progression in children with new-onset type 1 diabetes.

Research design and methods: A retrospective, longitudinal cohort study of children (<18 years) with type 1 diabetes (N=46) was included at diagnosis and followed until complete insulinopenia (C-peptide <0.03 nmol/L). Children were grouped into rapid progressors (n=20, loss within 30 months) and slow progressors (n=26). A sex-matched control group of healthy children (N=45) of similar age was included for comparison. Multiple biomarkers were assessed by proximity extension assay (PEA) at baseline and follow-up.

Results: At baseline, rapid progressors had lower C-peptide and higher autoantibody levels than slow. Three biomarkers were higher in the rapid group: carbonic anhydrase 9, corticosteroid 11-beta-dehydrogenase isozyme 1, and tumor necrosis factor receptor superfamily member 21. In a linear mixed model, 25 proteins changed over time, irrespective of group. One protein, a coxsackievirus B-adenovirus receptor (CAR) increased over time in rapid progressors. Eighty-one proteins differed between type 1 diabetes and healthy controls. Principal component analysis could not distinguish between rapid, slow, and healthy controls.

Conclusions: Despite differences in individual proteins, the combination of multiple biomarkers analyzed by PEA could not distinguish the rate of progression in children with new-onset type 1 diabetes. Only one marker was altered significantly when considering both time and group effects, namely CAR, which increased significantly over time in the rapid group. Nevertheless, we did find some markers that may be useful in predicting the decline of the C-peptide. Moreover, these could potentially be important for understanding type 1 diabetes pathogenesis.