Baseline OCT Biomarkers Associated with Visual Acuity in Diabetic Macular Edema

Topic To determine effect estimates and certainty of evidence for the associations between baseline OCT biomarkers and (1) patient visual acuity (VA) and (2) changes in VA from baseline to 6, 12, and 24 months after initiation of anti–vascular endothelial growth factor, steroid, or laser treatment for diabetic macular edema. Clinical Relevance Understanding the prognostic value conferred by biomarkers can help predict disease activity and inform treatment decisions. Methods This review was registered in the International Prospective Register of Systematic Reviews (identifier, CRD42023487798). Ovid MEDLINE, EMBASE, and CENTRAL databases were searched. Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) guidelines informed certainty of evidence. Results Twenty-eight reports from 27 studies evaluating 75 biomarkers were included. No biomarker with at least a moderate certainty was associated with improved VA or change in VA. Results are reported in Early Treatment Diabetic Retinopathy Study letters. Five biomarkers were associated with reduced VA at 2 or more time points with moderate certainty: (1) hyperreflective retinal foci (HRF; 6 months, –6.5 [95% confidence interval (CI), –10.4 to –2.6]; 12 months, –7.3 [95% CI, –11.6 to –3.0], (2) hyperreflective choroidal foci (HCF; 6 months, –7.3 [95% CI, –13.3 to –1.3]; 12 months, –7.5 [95% CI, –11.9 to –3.0], (3) disorganization of retinal inner layers (DRIL; 6 months, –6.0 [95% CI, –11.7 to –0.3]; 12 months, –7.3 [95% CI, –12.8 to –1.7], (4) disrupted ellipsoid zone (EZ) or external limiting membrane (ELM; 6 months, –9.7 [95% CI, –15.4 to –3.9]; 12 months, –7.5 [95% CI, –11.9 to –3.0]; 12 months isolated EZ disruption, –5.4; [95% CI, –9.2 to –1.6] ; 24 months, –9.0 [95% CI, –14.3 to –3.6], and (5) disrupted cone outer segment termination (COST) line (12 months, –8.5 [95% CI, –13.5 to –3.5]; 24 months, –8.8 [95% CI, –14.0 to –3.6]. Discussion Baseline HRF, HCF, DRIL, disrupted EZ or ELM, and disrupted COST lines were associated with worse VA at 2 or more time points with moderate certainty of evidence. Greater standardization in biomarker classification and better control of confounding variables are needed. Financial Disclosure(s) Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.