Measuring intraocular antibodies in eyes treated with anti-vascular endothelial growth factor

Objective: To characterize the total intraocular aqueous humour antibody profiles in cases receiving anti-vascular endothelial growth factor (anti-VEGF) for retinal vascular disease compared with controls without retinal pathology. Design: Cross-sectional. Participants: 93 aqueous humour samples: 22 eyes undergoing cataract surgery (controls) and 71 eyes receiving intravitreal injections (IVI) (cases) for macular edema or neovascularization. Methods: Antibody isotyping of aqueous humour was performed using Milliplex MAP Human Isotyping Multiplex Assay. Cases and controls were compared for several outcome measures. Results: The primary outcome measure was total mean antibody isotype concentration quantified in the aqueous humour. Secondary outcomes included comparing aqueous humour concentrations with visual acuity, number of IVI received, type of anti-VEGF agent injected, and persistence intra-/subretinal fluid post injection. Mean immunoglobulin M (IgM) concentrations in cases were 19-fold higher compared with controls. Aqueous immunoglobulin G (IgG)(1,2,3,4) and immunoglobulin A (IgA) were 2-4-fold higher in cases compared with controls. Disease-specific trends were observed, with diabetic retinopathy (DR) eyes containing the highest amounts of aqueous antibodies. Total number of injections correlated with higher titres of IgG(1) (p < 0.001), IgG(2) (p < 0.009), and IgG(3) (p < 0.001) in all cases analyzed with the strongest correlations seen in DR eyes (r = 0.77, p < 0.001). Presence of aqueous humour antibodies correlated with worse post-IVI best-corrected visual acuity; IgG(1) (p < 0.01), IgG(2) (p < 0.005), IgG(3) (p < 0.01), and IgA (p < 0.003) in all cases analyzed, with the strongest correlations seen in DR eyes (r = 0.74, p < 0.001). Conclusions: Intraocular antibodies are present in the aqueous humour at significantly higher concentrations in eyes receiving IVIs for retinal vascular diseases compared with controls.