Abstract
PURPOSE: To critically evaluate evidence supporting the angiogenesis-inflammation hypothesis in diabetic retinopathy (DR); to examine the roles of angiogenic growth factors and inflammatory cytokines in the biphasic effects of insulin on DR - early worsening followed by long-term risk reduction; and to explore broader implications of this hypothesis, including potential biphasic effects of semaglutide due to its insulinotropic action and strategies to mitigate the transient DR exacerbation associated with insulin and insulin secretagogues. METHODS: Literature review. RESULTS: Hyperglycemia and the accumulation of advanced glycation end products upregulate angiogenic growth factors and inflammatory cytokines, notably vascular endothelial growth factor (VEGF) and angiopoietin-2. These mediators compromise the retinal neurovascular unit and disrupt the blood-retinal barrier, while promoting endothelial adhesion molecule expression. The resulting leukostasis triggers hypoxia, leukocyte activation, and self-perpetuating ischemic-inflammatory loops. Insulin-like growth factor-1 further enhances retinal VEGF and angiopoietin-2 expression, potentially driving DR to the proliferative stage. Semaglutide, a glucagon-like peptide-1 (GLP-1) receptor agonist, has paradoxically been linked to DR deterioration like insulin therapy. In contrast, insulin-independent antidiabetic agents such as empagliflozin and metformin are not associated with DR worsening and may even slow its progression. GLP-1 augments glucose-induced insulin secretion, whereas insulin itself upregulates retinal VEGF and angiopoietin-2 and exerts pro-inflammatory effects in insulin-resistant states. Despite early worsening, long-term insulin therapy reduces DR risk through sustained glycemic control, with benefits outweighing initial harm. Semaglutide addresses a broader spectrum of systemic risk factors for DR than insulin, including hypertension, dyslipidemia, chronic kidney disease, and obesity. In addition, experimental evidence suggests that semaglutide may provide direct neurovascular protection through activation of the retinal GLP-1 receptor. CONCLUSIONS: The early worsening of DR associated with insulin therapy may be more accurately attributed to insulin’s pro-angiogenic effects rather than solely to rapid glycemic correction. As an insulin secretagogue, semaglutide may similarly cause early DR progression. However, long-term semaglutide therapy is anticipated to reduce DR risk by improving glycemic control, addressing a broader range of systemic risk factors compared with insulin, and offering direct neurovascular protection in the retina. Combination therapy with insulin-independent glucose-lowering agents may help mitigate the transient DR deterioration linked to insulin and insulin secretagogues. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s40942-026-00811-8.