Schizophrenia is associated with a reduced life expectancy and poor physical health, including neurovascular abnormalities, detectable in the retina, which is a direct extension of the brain. Optical coherence tomography (OCT) studies reveal that retinal thinning in patients with schizophrenia includes reduced nerve fiber layer and inner retinal atrophy. The observed retinal thinning may be confounded by disease-related factors such as medication use and disease duration. Remarkably, similar retinal changes appear in unaffected first-degree relatives, which indicates a genetic basis. Polygenic risk scores (PRS) are useful tools to investigate the connection between genetic variations and schizophrenia risk.
The present study, published in Nature Mental Health, aimed to investigate whether PRS for schizophrenia correlates with retinal morphology in healthy subjects. In this population study, 64,283 individuals were recruited with available retinal and genetic information from the UK Biobank. Individuals were excluded if they had non-British/Irish ancestry, eye disorders or hyperopia, quality control measures of genetic and OCT data, incomplete data related to body mass index (BMI), Townsend index, smoking, and alcohol drinker status (n = 29,344). Statistical analysis was performed using robust linear regression analysis and Pearson correlation coefficient analysis.
A total of 34,939 white Irish and British healthy individuals (mean age = 56.87±7.99 years, male = 45.42%, female = 54.58%, BMI = 27.23±4.67 Kg/m2) and 114 patients with schizophrenia were included in this study. Participants had overall macular thickness of 278.14±20.37 in the right eye and 275.69±20.37 in the left eye.
In robust regression analysis, a negative correlation was observed between PRS for schizophrenia and macular thickness with b = −0.17, 95% confidence interval (CI) of −0.31 to −0.03 and P = 0.018. A similar negative relation was observed between PRS for schizophrenia and outer retinal thickness with β = −0.10, 95% CI of −0.18 to −0.02, P = 0.018, and P value of family-wise error rate (pFWER) of 0.04.
Higher PRS related to neuroinflammation in schizophrenia were significantly correlated with thinner ganglion cell inner plexiform layers with b = −0.10, self-contained P = 0.014, and competitive P = 0.023. There were no statistically significant associations observed for other schizophrenia-related gene pathways like Gene Ontology Biological Process (GOBP) WNT signaling pathway, signal transduction (ST) WNT β-catenin pathway, Biocarta transforming growth factor-β (TGFβ) pathway, GOBP acute and chronic inflammatory microvascular pathways, and GOBP positive regulation of dopamine receptor signaling pathway.
The neuroinflammatory-specific PRS showed a significant indirect effect on the ganglion cell inner plexiform layer through C-reactive protein (CRP) of −0.001, with a 95% bootstrap CI of −0.003 to −0.0002 and P = 0.005. This indicated that the negative effect of risk on the retinal layer was partially mediated by elevated levels of CRP.
This study’s limitations include a non-representative UK Biobank due to a lower response rate, a younger population, and socioeconomic factors. It is limited to Irish/British populations. Robust linear analysis may dilute genuine extremes, and temporal data gaps could introduce confounding factors. Additionally, there was limited clinical significance due to the minimal retinal changes expected to affect visual acuity.
In conclusion, this study highlights the significant association between PRS for schizophrenia and specific retinal layer changes in undiagnosed individuals. This suggests the retina serves as a potential marker of genetic vulnerability for schizophrenia. Further research is needed to evaluate the interaction between genetics, environment, and comorbidities in order to assess the retinal changes as a potential biomarker for schizophrenia.
Reference: Rabe F, Smigielski L, Georgiadis F, et al. Genetic susceptibility to schizophrenia through neuroinflammatory pathways is associated with retinal thinness. Nat Mental Health. 2025. doi:10.1038/s44220-025-00414-6
