IGF1, serum glucose, and retinopathy of prematurity in extremely preterm infants

BACKGROUND Hyperglycemia, insulin insensitivity, and low IGF1 levels in extremely preterm infants are associated with an increased risk of retinopathy of prematurity (ROP), but the interactions are incompletely understood. METHODS In 117 extremely preterm infants, serum glucose levels and parenteral glucose intake were recoded daily in the first postnatal week. Serum IGF1 levels were measured weekly. Mice with oxygen-induced retinopathy alone versus oxygen-induced retinopathy plus streptozotocin-induced hyperglycemia/hypoinsulinemia were assessed for glucose, insulin, IGF1, IGFBP1, and IGFBP3 in blood and liver. Recombinant human IGF1 was injected to assess the effect on glucose and retinopathy. RESULTS The highest mean plasma glucose tertile of infants positively correlated with parenteral glucose intake [r(39) = 0.67, P < 0.0001]. IGF1 plasma levels were lower in the high tertile compared with those in low and intermediate tertiles at day 28 (P = 0.038 and P = 0.03). In high versus lower glucose tertiles, ROP was more prevalent (34 of 39 versus 19 of 39) and more severe (ROP stage 3 or higher; 71% versus 32%). In oxygen-induced retinopathy, hyperglycemia/hypoinsulinemia decreased liver IGF1 expression (P < 0.0001); rh-IGF1 treatment improved normal vascular regrowth (P = 0.027) and reduced neovascularization (P < 0.0001). CONCLUSION In extremely preterm infants, high early postnatal plasma glucose levels and signs of insulin insensitivity were associated with lower IGF1 levels and increased ROP severity. In a hyperglycemia retinopathy mouse model, decreased insulin signaling suppressed liver IGF1 production, lowered serum IGF1 levels, and increased neovascularization. IGF1 supplementation improved retinal revascularization and decreased pathological neovascularization. The data support IGF1 as a potential treatment for prevention of ROP. TRIAL REGISTRATION ClinicalTrials.gov NCT02760472 (Donna Mega). FUNDING This study has been supported by the Swedish Medical Research Council (14940, 4732, 20144-01-3, and 21144-01-3), a Swedish government grant (ALFGB2770), Lund medical faculty grants (ALFL, 11615 and 11601), the Skåne Council Foundation for Research and Development, the Linnéa and Josef Carlsson Foundation, the Knut and Alice Wallenberg Foundation, the NIH/National Eye Institute (EY022275, EY017017, EY017017-13S1, and P01 HD18655), European Commission FP7 project 305485 PREVENT-ROP, Deutsche Forschungsgemeinschaft (CA-1940/1-1), and Stiftelsen De Blindas Vänner.