A longitudinal study of serum insulin-like growth factor-I levels over 6 years in a large cohort of children and adolescents with type 1 diabetes mellitus: A marker reflecting diabetic retinopathy

OBJECTIVE: To evaluate longitudinal serum insulin-like growth factor-I (IGF-I) in a large cohort of children and adolescents with type 1 diabetes in relation to hemoglobin A1c (HbA1c), age, diabetes duration, and body mass index (BMI), its association to height and retinopathy, and in comparison with healthy subject references.

METHODS: A total of 2683 serum IGF-I values were obtained from 806 children and adolescents with T1DM, from annual blood samples for up to 6 consecutive years.

RESULTS: In a multiple regression analysis IGF-I values were negatively correlated to HbA1c and diabetes duration, and positively correlated to BMI (P < .001, P < .001, and P < .001, respectively, adjusted r2  = 0.102). Children and adolescents with T1DM had lower mean IGF-I levels and reference interval limits compared to healthy subjects. In boys, mean (SD) IGF-I SD score (SDS) levels were -1.04 (±1.3) calculated from the healthy reference. IGF-I peaked at 15 years of age, similarly to healthy controls, but with markedly lower levels in late puberty. Girls were more affected at later stages of puberty but with a slightly less depressed overall mean IGF-I SDS of -0.69 (±1.2). In a subgroup of 746 subjects with fundus photography, a negative correlation was seen between individual mean IGF-I SDS and preproliferative retinopathy (P = .004, adjusted r2  = 0.021). In another subgroup of 84 adolescents, no correlation was seen between individual mean IGF-I SDS and target height SDS or distance to target height SDS.

CONCLUSION: Poor metabolic control and diabetes duration impact negatively on serum IGF-I levels. A low individual mean IGF-I level was associated with progression of retinopathy independently of HbA1c, age, gender, and diabetes duration. Disease, sex and age related IGF-I SDS may become clinical helpful as a supplement to HbA1c in predicting the long-term outcome for children and adolescents with T1DM.