Retinal neurodegeneration may be detectable in young people with type 1 diabetes before clinically apparent diabetic retinal disease emerges, according to a prospective cohort study published in JAMA Ophthalmology.
The study, led by researchers at Johns Hopkins School of Medicine, Baltimore, examined whether retinal layer thickness on optical coherence tomography (OCT) was associated with glycemic outcomes and diabetic retinal disease (DRD) in youth with type 1 diabetes. Although DRD is typically defined by vascular changes on fundus imaging, growing evidence suggests that neuroretinal damage may occur earlier. In adults, retinal layer thinning has been linked to diabetes and DRD, but pediatric data – particularly in youth with type 1 diabetes of substantial duration – have remained limited.
The investigators analyzed participants from ACCESS2, the AI for Pediatric Diabetic Eye Exams Study 2, conducted at the Johns Hopkins Pediatric Diabetes Center. The cohort included 294 young people aged 9 to 21 years with type 1 diabetes, contributing 578 interpretable eyes. Mean age was 15.8 years, median diabetes duration was 7.0 years, and median hemoglobin A1C (HbA1C) was 8.5 percent. Most participants were using diabetes technology: 87.4 percent were prescribed continuous glucose monitoring and 71.4 percent used an insulin pump.
Participants underwent fundus photography and spectral-domain OCT imaging. OCT volumes were segmented to assess retinal nerve fiber layer (RNFL) thickness, ganglion cell and inner plexiform layer (GCL+IPL) thickness, combined GCL+IPL+RNFL thickness, total retinal thickness, and outer retinal layer thickness. Fundus images were graded by the Wisconsin Reading Center using the International Clinical Diabetic Retinopathy Scale.
Clinically apparent DRD was uncommon but present. Among 578 eyes, 65 eyes had mild DRD and 10 had moderate DRD. At the participant level, 77.8 percent had no apparent retinopathy, 18.8 percent had mild nonproliferative DRD, and 3.4 percent had moderate DRD.
The key finding was that retinal thinning tracked more clearly with glycemic exposure than with retinopathy status. In adjusted analyses, the presence or severity of DRD was not significantly associated with total retinal thickness, RNFL thickness, GCL+IPL thickness, or outer retinal layer thickness. By contrast, higher HbA1C was associated with thinner GCL+IPL and outer retinal layers. Each 1 percentage point increase in HbA1C was linked to a 0.39 μm lower GCL+IPL thickness and a 0.81 μm lower outer retinal layer thickness. Among participants using continuous glucose monitoring, greater time in the target glucose range of 70–180 mg/dL was associated with greater outer retinal layer thickness.
The study authors note that these findings support the concept of diabetic retinal neurodegeneration as an early process, potentially preceding visible vascular disease. This observation adds weight to the idea that OCT may eventually help identify young patients with type 1 diabetes who are undergoing subclinical retinal change before standard retinopathy findings appear.
The findings are exploratory and not yet practice-changing. The study was conducted at a single institution, included relatively few eyes with moderate DRD, and lacked longitudinal follow-up. Still, it offers one of the largest pediatric OCT datasets in type 1 diabetes to date. The next step: long-term studies must determine whether early OCT thinning predicts later diabetic retinal disease – and whether improved glycemic control can preserve the young neuroretina.
